Innuos Zenith Mk3 or OpticalRendu + separate server?

[BigAlMc]

2 minutes ago, vortecjr said:

However, I don't think they matter as used with our Rendu series

 

Sorry dude but need to disagree. The Zenith has USB out and a second LAN port for Ethernet out. My Directstream DAC supports both USB and Ethernet inputs.

 

Amongst other things I compared:

 

Server to MicroRendu to DAC

 

vs

 

Zenith via Ethernet output to DAC

 

The Zenith Ethernet sounded better. Others have reported similarly.

 

Ethernet might offer inbuilt isolation but its not the whole story. What you input into that Ethernet cable matters too. 

 

We can agree to disagree if you want.

 

9 minutes ago, vortecjr said:

Martin who is a great industry member and a top Sonore dealer and support center.  

 

Agreed. Great bloke and very helpful.

 

Cheers,

Alan

[vortecjr]

I know what thread you have been hanging out in...LOL. I even have a strong opinion about this connection scheme, but no worries we will be addressing that connection scheme very soon;)

 

I have to take your perception on face value though because I have multiple customers on all sides of this argument. Some like the USB input, some like the LVDS i2s input, some like the Ethernet input, and some can't decide. Each one has it's merits (pro and con) and its really hard to compare them. It can be anything that drives the opinions one way or the other. I don't try to understand why or convince people one is better than the other even though we know this DAC very well.  

[vortecjr]

In one of posts above you said this referring to Barrows, "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." This is an opinion or conclusion formed on the basis of incomplete information. I'm sure some have and you may have. However, I have access to the back end data from the usage of www.sonicorbiter.com and the data suggests otherwise. If I just take a multi day snap shot of the units online for the microRendu product its usage is actually quite high. Another indicator is the number of people who have updated their units to the most current version of SonicOrbiter. These indicators and more suggest that the microRendu is still being enjoyed by many no matter what some vocal few suggest.   

[matthias]

49 minutes ago, BigAlMc said:

The Zenith Ethernet sounded better. Others have reported similarly.

 

Yes, in his review of the Innuos Statement Jason Kennedy reported the same.

Further he found that the SQ was even higher when the Statement was used as a pure server with ethernet out into either a Naim or Auralic G2 streamer.

 

Matt

[BigAlMc]

1 hour ago, vortecjr said:

In one of posts above you said this referring to Barrows, "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." This is an opinion or conclusion formed on the basis of incomplete information. I'm sure some have and you may have. However, I have access to the back end data from the usage of www.sonicorbiter.com and the data suggests otherwise. If I just take a multi day snap shot of the units online for the microRendu product its usage is actually quite high. Another indicator is the number of people who have updated their units to the most current version of SonicOrbiter. These indicators and more suggest that the microRendu is still being enjoyed by many no matter what some vocal few suggest.   

 

People sell the MicroRendu when they move on. I did

The second hand demand is pretty strong which is a complimentary reflection on your product.

 

But the number of MicroRendus still in use doesn't rule out that many people moved onwards to something else in search of higher SQ or some other functionality.

[vortecjr]

32 minutes ago, BigAlMc said:

 

People sell the MicroRendu when they move on. I did

The second hand demand is pretty strong which is a complimentary reflection on your product.

 

But the number of MicroRendus still in use doesn't rule out that many people moved onwards to something else in search of higher SQ or some other functionality.

I could say the same thing about any product...including yours. People here have reporting moving on from it to a NUC:) Howerver, I would not use that fact as an argument about the products relavency because it means absolutely nothing.

[McNulty]

@BigAlMc: if I understand correctly, you use the Zenith with its ethernet output. Would you say there's a large SQ delta with the Zeniths USB output directly connected with the DAC?

[BigAlMc]

Hi @McNulty,

 

No, the argument I was making was that the 'clean' Ethernet output from the Zenith was superior to the output from the MicroRendu (in my experience at least)

 

The Ethernet and USB outputs from the Zenith are pretty well matched.

 

FWIW I ended up using USB output far more but that was due to other reasons such as the LMS player only using USB, and the contributions of other USB peripherals such as the Lush^2 cable and the TX-USBUltra.

 

My debate with the gents from Sonore  here is that they claim that the isolation offered by Ethernet and now Optical when coupled with their low noise, designed for audio, Rendu board means the server is irrelevant. I disagree because my experience is that its better to not introduce that noise in the first place, using a good server like the Zenith. Rather than to rely on removing it later.

 

I think you should go for the Zenith. But as per my original post I can't say for sure that the Optical Rendu wouldn't outperform it.

 

Also my apologies for hijacking your thread!

 

Cheers,

Alan

[Blackmorec]

I agree 100% with Alan’s post. The USB output of the Zenith direct into my DAC provided the absolute best sound I’ve every heard from digital. I also agree that avoiding noise in the first place is superior to trying to remove it later on.  Why do I say that? I use a wireless connection between my router and server which should fully isolate the hi-fi, yet everything done at the modem/server level, for example ps improvements clearly reflects in the sound quality. Its as I said in a previous post...theory doesn’t always match practice in digital audio when the criteria is sound quality

[matthias]

1 hour ago, Blackmorec said:

The USB output of the Zenith direct into my DAC provided the absolute best sound I’ve every heard from digital.

 

Did you compare USB output to Ethernet output of your Statement into your Devialet?

Thanks

 

Matt

[Blackmorec]

23 minutes ago, matthias said:

 

Did you compare USB output to Ethernet output of your Statement into your Devialet?

Thanks

 

Matt

No, I haven’t done that yet. If you look at the interfaces on the Statement, both USB and Ethernet have OCXO 3ppb clocks and dedicated LPS rails so much of the differences reported could be down to cables and interfaces on the DAC side. 

 

I’m just waiting for for my last DC3 PS to run in....once it has I’ll give an ethernet cable a try and report back. I have a spare Meicord lying around so while its not quite up to the SR standard, it isn’t chopped liver either. 

[barrows]

On 2/23/2019 at 2:24 AM, BigAlMc said:

My debate with the gents from Sonore  here is that they claim that the isolation offered by Ethernet and now Optical when coupled with their low noise, designed for audio, Rendu board means the server is irrelevant. I disagree because my experience is that its better to not introduce that noise in the first place, using a good server like the Zenith. Rather than to rely on removing it later.

My only comments in this thread refer to the opticalRendu and Signature Rendu optical.  Both products which you could not possibly have any experience with.  While wired (say CAT 6A) Ethernet connections do allow some noise to couple through ( while still providing a great deal of isolation), optical Ethernet connection is a different thing entirely.  When one uses an optical Ethernet connection one does not "introduce" any noise to downstream components, any noise from the upstream server, router, switches, modems, whatever, does not travel on the optical connection.

 

Additionally,  Rendu products are not deigned to "remove" noise, indeed they are created to be as noise free as possible intrinsically.  And optical Ethernet connections do not "remove" noise either, they just cannot transmit electrical noise from the server (be it a NAS, a standard commercial computer, or an "audiophile" server) to the Renderer.

[Summit]

I don’t understand why the Sonore reps are allowed to promote their gear and tech in every thread. 

[matthias]

2 hours ago, barrows said:

My only comments in this thread refer to the opticalRendu and Signature Rendu optical.  Both products which you could not possibly have any experience with.  While wired (say CAT 6A) Ethernet connections do allow some noise to couple through ( while still providing a great deal of isolation), optical Ethernet connection is a different thing entirely.  When one uses an optical Ethernet connection one does not "introduce" any noise to downstream components, any noise from the upstream server, router, switches, modems, whatever, does not travel on the optical connection.

 

Basically it is the same with WiFi, the conversion from one medium to another and back. Noise is blocked in both cases. In both cases the signal integrity depends on how perfect these double conversions are performed. When these conversions are perfect, one gets the same signal minus the blocked noise as before the double conversion. 

In both cases a better source upstream will generate a better signal. The better signal will be present after the double conversion. Period.

 

Matt

[barrows]

3 hours ago, matthias said:

better source upstream will generate a better signal

I am curious as to what the above means: what constitutes this "better signal" which you speak of?  What do you think is being lost by a conversion which is not "better"?

And how does this "better signal" result in a sound quality improvement, as here I am making the assumption that this is what are implying, right?

 

Additionally: what are the implications for "signals" streamed from the Internet, such as Tidal and Qobuz?  Certainly if this "signal", "problem" is actually a real thing which results in sound quality degradation, the fact that Tidal and Qobuz are not using "audiophile" servers to send their signals must be a big problem, right, especially considering that their "signals" are traveling through many different devices on the way to our homes.

[barrows]

6 hours ago, Summit said:

I don’t understand why the Sonore reps are allowed to promote their gear and tech in every thread. 

The OP asked questions specifically about Sonore gear-is there some reason you think we should not share our experiences and knowledge?  I would suggest also, as Innouous was mentioned by the OP, I would welcome representatives from that company to respond here as well, and contribute to the discussion.

[McNulty]

As the OP I have no problem with the Sonore guys explaining their view on the best streaming approach and the implementation in their upcoming product. 

[matthias]

4 hours ago, barrows said:

I am curious as to what the above means: what constitutes this "better signal" which you speak of?  What do you think is being lost by a conversion which is not "better"?

And how does this "better signal" result in a sound quality improvement, as here I am making the assumption that this is what are implying, right?

Additionally: what are the implications for "signals" streamed from the Internet, such as Tidal and Qobuz?  Certainly if this "signal", "problem" is actually a real thing which results in sound quality degradation, the fact that Tidal and Qobuz are not using "audiophile" servers to send their signals must be a big problem, right, especially considering that their "signals" are traveling through many different devices on the way to our homes.

 

Let me suggest to read this thread for this topic:

https://www.whatsbestforum.com/threads/taiko-audio-sgm-extreme-the-crème-de-la-crème.27433/

 

I hope I can listen to it at HIGH END 2019 in Munich.

 

This post might be very interesting for you too:

https://www.whatsbestforum.com/threads/taiko-audio-sgm-extreme-the-crème-de-la-crème.27433/page-6#post-557413

 

BTW, Taiko Audio reported that in the Netherlands streaming from Tidal sounds better than local files from a NAS while in the US it is the contrary.......

 

Matt

[barrows]

3 hours ago, matthias said:

BTW, Taiko Audio reported that in the Netherlands streaming from Tidal sounds better than local files from a NAS while in the US it is the contrary.......

I guess a listener would have to be in both places at the same time to make that comparison. 🤣

[barrows]

3 hours ago, matthias said:

 

Let me suggest to read this thread for this topic:

https://www.whatsbestforum.com/threads/taiko-audio-sgm-extreme-the-crème-de-la-crème.27433/

 

I hope I can listen to it at HIGH END 2019 in Munich.

 

This post might be very interesting for you too:

https://www.whatsbestforum.com/threads/taiko-audio-sgm-extreme-the-crème-de-la-crème.27433/page-6#post-557413

 

BTW, Taiko Audio reported that in the Netherlands streaming from Tidal sounds better than local files from a NAS while in the US it is the contrary.......

 

Matt

There is no information in those threads which answers any of my questions.  Just folks throwing **it at the wall to see what sticks.  I am well aware of much of the development for many "audiophile" servers, maybe you are not aware that Sonore used to build custom servers before we realized that better performance was achievable for way less expense via Ethernet audio distribution.

 

Again:  I am looking for your answer to this question:

 

I am curious as to what the above means: what constitutes this "better signal" which you speak of?  What do you think is being lost by a conversion which is not "better"?

And how does this "better signal" result in a sound quality improvement, as here I am making the assumption that this is what are implying, right?

 

[Blackmorec]

23 hours ago, barrows said:

My only comments in this thread refer to the opticalRendu and Signature Rendu optical.  Both products which you could not possibly have any experience with.  While wired (say CAT 6A) Ethernet connections do allow some noise to couple through ( while still providing a great deal of isolation), optical Ethernet connection is a different thing entirely.  When one uses an optical Ethernet connection one does not "introduce" any noise to downstream components, any noise from the upstream server, router, switches, modems, whatever, does not travel on the optical connection.

 

Additionally,  Rendu products are not deigned to "remove" noise, indeed they are created to be as noise free as possible intrinsically.  And optical Ethernet connections do not "remove" noise either, they just cannot transmit electrical noise from the server (be it a NAS, a standard commercial computer, or an "audiophile" server) to the Renderer.

Given that we are comparing actual products (Servers) with future Sonore products which are currently unavailable and therefore not tried by anyone, the entire discussion is theoretical and fairly pointless. As I’ve mentioned in previous posts, in digital hardware the actual performance doesn’t always follow what’s predicted in theory so its always important to listen to products before buying.  During the building of my current system I tried a number of networking and isolation strategies and the very best I found was wi-fi based on LPS DC supplies throughout, feeding an Innuos Server to my Devialet 440s.  Despite the total electrical isolation achieved, several other steps like vibration isolation and selection of both upstream and downstream cables made a difference, so the conclusion I would reach is that complete isolation, while helpful, isn’t the entire solution. I also reclock the data stream before the server, yet ameliorating vibration still improved sound quality. 

Going back to the OPs question,  an Innuos Zenith server fed by an optimised wi-fi stream is in my experience and my installation the best sound quality I could achieve while being extremely straightforward to install and robust in operation

 

[barrows]

6 hours ago, Blackmorec said:

Given that we are comparing actual products (Servers) with future Sonore products which are currently unavailable and therefore not tried by anyone

Actually, these products have been tried by both Chris Connaker and the people who work with Sonore, like me.  This is exactly why I am commenting here.

WiFi has a few of inherent problems: one is that it generally does not achieve the same bandwidth as wired (or optical) connections, meaning there will be more data errors and hence more re-sends to correct those errors (please note I am not claiming there will be uncorrected errors).  It is plausible, but unproven, that more error correction by the endpoint will generate more processor noise, and could result in a decrease in sound quality.  In any case, for that reason alone I advise folks to avoid WiFi for music distribution if possible.  Another problem is that while WiFi systems should have enough bandwidth and signal integrity to transmit high resolution audio without hiccups, in practice, many Sonore customers have reported problems (dropouts, etc).  The other problem is the WiFi transceiver itself, this is by definition a source of noise, as it is a radio signal producer, and putting an actual radio transmitter in close proximity to the audio system is asking for trouble: this is why Sonore Renderers do not offer WiFi as an option built in (and why dCS who does offer WiFi advises not to use WiFi for best sound quality).

Now if you get great results with WiFi, good, I am happy for you.  Given the inherent potential problems with WiFi, I would not recommend that approach.  But, there is always more than one way to skin a cat, and if it works for you I am fine with that.

 

I will give a simple example of a test where I found upstream changes to make no difference in an optically connected system. I use a Mac Mini as a server.  @Superdad, has noted the the built in power supply of the Mini has a very large leakage current  problem.  So, I ground my Mini to shunt the leakage current and make it a non issue.  With wired Ethernet, this did result in a small improvement in sonics, with an optically connected system I cannot hear any difference with the Mini grounded or not (this is very easy to A/B as the ground connection is via a plug).  I have a way I can measure some of the electrical noise components coming over wired Ethernet to the audio system.  For example, with a wired (CAT 6A) set up, I can measure noise, even on the AC cables of each audio component.  With an optically cabled Ethernet connection, that noise is entirely gone.

 

Upstream electrical noise is fully isolated by the optical Ethernet connection.  Now, where there is room to investigate further (which I have mentioned in this thread) is with the performance of the upstream clocks.  While it is virtually impossible to come up with a viable explanation as to how these clocks would matter, considering the asynchronous transmission of Ethernet data, and the fact that the Internet actually works, despite the loads of accumulating phase noise from thousands of clocks along the way (if indeed that inaccuracy actually accumulates) there is enough anecdotal information to warrant further investigation.  So John Swenson is investigating this reported phenomena, and is trying to measure it-I look forward to seeing his results, and perhaps an explanation-because right now, no one has ever offered even the wildest theory as to how this could matter.  In any case, if the upstream clock's performance really is an issue, this could explain why you report that vibration control of your upstream components results in improvements in your WiFi isolated system, as clocks at these levels of phase noise are quite sensitive to vibration.  When the Sonore optical module is available, it will be easy to test (anecdotally, by listening at least) whether the upstream clock(s) actually matter, as it will have a femto level clock with a clean power rail: I look forward to making this test. 

[Superdad]

5 minutes ago, barrows said:

 @Superdad, has noted the the built in power supply of the Mini has a very large leakage current  problem.  So, I ground my Mini to shunt the leakage current and make it a non issue. 

 

Just FYI: 

Grounding the computer only shunts the high-source-impedance leakage; It does nothing to eliminate the much larger amplitude, broadband, low-source-impedance AC leakage inherent in all SMPS units.  And our course the Mac mini is far from the only computer with an SMPS.

 

5 minutes ago, barrows said:

Now, where there is room to investigate further (which I have mentioned in this thread) is with the performance of the upstream clocks.  While it is virtually impossible to come up with a viable explanation as to how these clocks would matter, considering the asynchronous transmission of Ethernet data, and the fact that the Internet actually works, despite the loads of accumulating phase noise from thousands of clocks along the way (if indeed that inaccuracy actually accumulates) there is enough anecdotal information to warrant further investigation.  So John Swenson is investigating this reported phenomena, and is trying to measure it-I look forward to seeing his results, and perhaps an explanation-because right now, no one has ever offered even the wildest theory as to how this could matter. 

 

Actually, over a year ago John put forth a very viable and not wild theory as to the mechanisms:

https://audiophilestyle.com/forums/topic/37418-usb-jitter-why-does-it-matter/?do=findComment&comment=735512

 

And development of his sophisticated measurement system (which UpTone has been funding) is coming along well, though as you know, product development keeps getting in the way.  When it is producing results, it will demonstrate the difference between simple Ethernet galvanic isolation (ala an FMC such as the opticalModule) and true isolated clock domain reclocking that the EtherREGEN will provide.  

And while it is anyone's guess just yet as to how much difference will be heard, the widespread anecdotal reports of what folks have been hearing with various clocking (of Ethernet and USB) lead us to believe that our first of its kind isolated final clock domain solution (EtherREGEN) will be quite effective.

 

Exciting times! 

 

 

 

[barrows]

28 minutes ago, Superdad said:

Grounding the computer only shunts the high-source-impedance leakage; It does nothing to eliminate the much larger amplitude, broadband, low-source-impedance AC leakage inherent in all SMPS units.  And our course the Mac mini is far from the only computer with an SMPS.

@Superdad.  But the above noise does not, generally speaking, go through ethernet transformers, and is not an issue in an Ethernet connected system as per my example.  While I would be concerned about this noise in a directly connected computer: mini-USB-DAC, I would not in the Ethernet environment of my example.  I know you are trying to sell your Mini mods and the very nice JS-2 supply here...

I need to get back to work here, lots on my plate, but will read JS' hypothesis linked above when I have a moment.  BTW, that title is USB jitter, and not what we are talking about here, USB jitter is well addressed by the Femto clocking in the Rendu's.  But perhaps the thesis addresses ethernet as well...

[Superdad]

5 minutes ago, barrows said:

@Superdad.  But the above noise does not, generally speaking, go through ethernet transformers, and is not an issue in an Ethernet connected system as per my example.  While I would be concerned about this noise in a directly connected computer: mini-USB-DAC, I would not in the Ethernet environment of my example.

 

Ah, yes, you are correct in that case. 

 

5 minutes ago, barrows said:

I know you are trying to sell your Mini mods and the very nice JS-2 supply here...

 

Very funny.  That was not the point of my post.  Beside, we are oversold on JS-2s nearly every month.  Can't build more than 25 per month. 

 

5 minutes ago, barrows said:

I need to get back to work here, lots on my plate, but will read JS' hypothesis linked above when I have a moment.  BTW, that title is USB jitter, and not what we are talking about here, USB jitter is well addressed by the Femto clocking in the Rendu's.  But perhaps the thesis addresses ethernet as well...

 

The hypothesis applies to Ethernet as well.  You'll see that John indicates that.  And it is why we are going to the trouble and expense to use 600MHz active digital isolators, differential clocking, and 10GHz differential flip-flops--all in their own separate clock and power domain--in the EtherREGEN.