Innuos Zenith Mk3 or OpticalRendu + separate server?

[McNulty]

Hi guys, this is my first post after lurking the forum for a couple of weeks, hope you can help me a bit in my quest for a new streamer.

 

As a bit of background, I intend to use the streamer with Roon. My DAC can handle up to DSD128 and I will only use one Roon streamer/endpoint. The full DIY route (AL/NUC) is not for me, I'm looking for a more off the shelf solution.

 

As I already have an optical network in place, I was leaning towards the Sonore OpticalRendu, powered by an LPS 1.2. This would mean I can get the FMC out of my system and connect the OpticalRendu directly to my Cisco switch, which has an optical connection. Apart from the OpticalRendu I would also need a Roon server such as a NUC with RoonRock (bit DIY) or a SonicTransporter (more off the shelf).  

 

As an alternative, the Innuos Zenith Mk3 seemed like a more simple one-box solution, and a separate server and an LPS for both the streamer and the server wouldn't be necessary. 

 

Roon themselves seem to advice a two-box solution, but on the other hand many users seem happy with the Innuos SQ and UX. I'm currently leaning towards the Innuos as the KISS option. On the other hand, the OpticalRendu and server could provide more flexibility for future upgrades. 

 

Dealer price for the Zenith Mk3 in Europe is EUR 3,395. Total price for the OpticalRendu / server / LPSs package would not be much lower I expect, taking into account transport and customs from US to Europe. 

 

Maybe this seems like an apples and oranges comparison, but I was still hoping maybe others have done a similar comparison. Not strictly limited to the OpticalRendu, as the model is not for sale yet, but a similar two-box vs one-box comparison. Note I've already booked a demo for the Zenith at a local dealer, but expect a demo of the OpticalRendu will not be possible. Thanks for any input you can provide. 

 

 

 

 

 

 

 

 

[jaspal kallar]

Perhaps take a look at this video:

https://darko.audio/2019/02/a-short-film-about-the-innuos-zen-mk3/

It's not the Zenith MK3 but the cheaper Zen Mk3.  

The reviewer who has the much acclaimed Zenith SE, thinks the Zen Mk3 comes quite close to the sound of the Zenith SE.

 

[BigAlMc]

Hi @McNulty,

 

Welcome. And loved your work in The Wire 😁

 

I can't answer your question as no one knows how the optical rendu will perform.

 

I can offer two small items of advice.

 

Firstly, I owned the Zenith MKII and still own the Zenith SE and both are superb. I'd heartily recommend Innuos.

 

Secondly if you're considering the SonicTransporter then take a look at the Audiostore Prestige from Vortexbox. 

 

https://www.vortexbox.co.uk

 

Martin Smith does great turnkey servers, offers superb service and is based in the UK. Nothing against Andrew Gillis that does the Sonictransporter who is similarly superb. But given you mention importing to Europe then Martin is based in the UK, offers essentially the same products and you'd avoid the hassle of customs charges etc from US to Europe. 

 

Martin also stocks Uptone Audio and Sonore.

 

Hope this helps a bit.

 

Cheers,

Alan

 

[k-man]

I’d recommend the OpticalRendu based on your existing optical network.

I would only recommend the Innuos over the OpticalRendu if you were needing the ripping functionality.

As for the server, I would suggest waiting for an optical based solution, or at least a short term cost effective solution i.e ROCK.

[barrows]

Consider, if you use an optical Network connection to a Renderer (or DAC with an optical input), there really is no reason to purchase an expensive "audiophile" server, as the optical connection isolates the Renderer from noise.  Just make sure you place the server (or NAS, whatever) away for the audio system and plugged into a different AC line form the audio system.

[BigAlMc]

4 hours ago, barrows said:

Consider, if you use an optical Network connection to a Renderer (or DAC with an optical input), there really is no reason to purchase an expensive "audiophile" server, as the optical connection isolates the Renderer from noise.  Just make sure you place the server (or NAS, whatever) away for the audio system and plugged into a different AC line form the audio system.

 

@barrows you should state your affiliation here because given you work for Sonore your advice isn't exactly impartial.

 

@McNulty,

 

What do you currently use as a server? And what inputs does your DAC accept and does it prefer any particular input?

 

Cheers,

Alan

[matthias]

5 hours ago, barrows said:

Consider, if you use an optical Network connection to a Renderer (or DAC with an optical input), there really is no reason to purchase an expensive "audiophile" server, as the optical connection isolates the Renderer from noise.  Just make sure you place the server (or NAS, whatever) away for the audio system and plugged into a different AC line form the audio system.

 

Basically wrong in my view.

With a better server you get a better signal and less noise. It is right that with something like an opticalRendu you have even less noise but the quality of the incoming signal is determined by the server. So the quality of the server is always important. Source first rules.

 

Matt

[Blackmorec]

Roughly a year ago I set out to establish a system for local and remote streaming from CD rips and  Quboz.  I intended to use Roon as I valued its ability to manage libraries and find new music. After a lot of evaluation I decided on the Zenith Mk2 but the SE was announced just before I ordered so that’s what I bought. 

The system arrived and within a few hours was producing magic, basically the best sound quality I’d ever heard. The installation was entirely straightforward, the Roon integration is absolutely solid and the OS is robust and very easy to use. 

 

The Zenith’s power supply is a major contributor to its great sound and the designers have paid a lot of attention to EMI and vibration control so both noise and jitter seem to be very low.

 

I found that plugging the Zenith into my system’s dedicated mains supply benefitted SQ, but plugging in another device’s SMPS had a seriously deleterious effect so I concluded that the Zenith was injecting very little noise to disturb the D to A conversion. 

 

The Quboz integration is seamless and response times virtually as fast as a local GUI when network  performance is adequate. 

 

In the year I’ve been using the Zenith the only single issue I’ve had was the loss of the entire internet connection during an electrical storm, otherwise not a single glitch

 

[matthias]

1 hour ago, Blackmorec said:

Roughly a year ago I set out to establish a system for local and remote streaming from CD rips and  Quboz.  I intended to use Roon as I valued its ability to manage libraries and find new music. After a lot of evaluation I decided on the Zenith Mk2 but the SE was announced just before I ordered so that’s what I bought. 

The system arrived and within a few hours was producing magic, basically the best sound quality I’d ever heard. The installation was entirely straightforward, the Roon integration is absolutely solid and the OS is robust and very easy to use. 

 

The Zenith’s power supply is a major contributor to its great sound and the designers have paid a lot of attention to EMI and vibration control so both noise and jitter seem to be very low.

 

I found that plugging the Zenith into my system’s dedicated mains supply benefitted SQ, but plugging in another device’s SMPS had a seriously deleterious effect so I concluded that the Zenith was injecting very little noise to disturb the D to A conversion. 

 

The Quboz integration is seamless and response times virtually as fast as a local GUI when network  performance is adequate. 

 

In the year I’ve been using the Zenith the only single issue I’ve had was the loss of the entire internet connection during an electrical storm, otherwise not a single glitch

 

 

What happened musically going from the SE to the Statement?

Thanks

 

Matt

[matthias]

33 minutes ago, matthias said:

 

What happened musically going from the SE to the Statement?

Thanks

 

Matt

@Blackmorec

I have already seen your answer here:

 

https://www.whatsbestforum.com/threads/a-new-star-in-2019-innuos-statement-server.27184/page-12#post-558663

 

Thanks

 

Matt

 

[barrows]

9 hours ago, BigAlMc said:

@barrows you should state your affiliation here because given you work for Sonore your advice isn't exactly impartial.

It is clearly stated in my Signature.  

 

My advice here was general, and completely unrelated to Sonore.  Also the original topic does refer to, and ask about, Sonore product.

[BigAlMc]

9 minutes ago, barrows said:

It is clearly stated in my Signature.  

 

My advice here was general, and completely unrelated to Sonore.  Also the original topic does refer to, and ask about, Sonore product.

 

Fair enough. I read and responded to your response on my phone and it's not obvious there as signatures are not shown.

 

I stand by my impartiality comment though. My experience is that everything matters. So whilst I hope the Optical Rendu is a great product and am a fan of Sonore, I'm not convinced that you can state with any certainty that it renders (pardon the pun) the server as irrelevant.

 

I think the Op would be pleased with the Innuos and it meets his needs. But whether the Optical Rendu would outperform it is impossible to say at this stage.

 

Apologies if my original comment was too harsh.

 

Cheers,

Alan

[barrows]

9 hours ago, matthias said:

Basically wrong in my view.

With a better server you get a better signal and less noise. It is right that with something like an opticalRendu you have even less noise but the quality of the incoming signal is determined by the server. So the quality of the server is always important. Source first rules.

Perhaps wrong in your view, but did you try it?

The entire reason to go to an optical Network solution is that it provides complete isolation from the upstream gear.  The "quality" of the upstream signal does not matter (unless it is so bad that you have lost bits, which does not happen on Ethernet transmissions).  An optical connection does not transmit ANY noise from the upstream gear to the endpoint, it cannot.

What does matter is the quality of the signal as it is presented to the DAC, and in this case, yes, the source does matter, but the source is the Renderer, not the computer on the other side of the optical isolation.

This is the entire reason to go with an optical Ethernet based solution.  You can run any type of computer to serve the files, and it will not matter one bit.  Just make sure that upstream computer gear is well isolated on the AC side from the audio system (plug it into a different circuit from the audio system, preferably on the opposite phase).

[barrows]

3 minutes ago, BigAlMc said:

I stand by my impartiality comment though. My experience is that everything matters. So whilst I hope the Optical Rendu is a great product and am a fan of Sonore, I'm not convinced that you can state with any certainty that it renders (pardon the pun) the server as irrelevant.

Even though the OP mentioned Sonore product specifically, I tried to keep my comments general RE optical connections, as there are other optical solutions available (like the Lumin's X-1, for example).

The fact is that optical isolation is absolute, there is no middle ground here, as long as the upstream gear is far away enough to not interfere with the audio system from airborne RFI, you will have no issues, no interference will come over the optical cable, it just cannot.

Anyone can test this with a couple of FMCs and a single optical cable.  Try any audiophile server on the other end, and then try a basic laptop.  See if you hear a difference...

 

While opinions are nice, there are some actual facts here-optical connections work so well because they are not subject to electrical interference the way electrical wiring is, this just a fact.  It is why optical is the choice for long distance, high rate data transmission.

 

I do not stream from Internet based music sources (Tidal, Qobuz, etc), but for those who do: do you suspect the server(s) these companies are using is an audiophile approved one?  Would Tidal sound better is their server(s) were Innuous?  This is, perhaps, a topic for another thread, but I am just offering it up as food for thought.

[BigAlMc]

36 minutes ago, barrows said:

Even though the OP mentioned Sonore product specifically, I tried to keep my comments general RE optical connections, as there are other optical solutions available (like the Lumin's X-1, for example).

The fact is that optical isolation is absolute, there is no middle ground here, as long as the upstream gear is far away enough to not interfere with the audio system from airborne RFI, you will have no issues, no interference will come over the optical cable, it just cannot.

Anyone can test this with a couple of FMCs and a single optical cable.  Try any audiophile server on the other end, and then try a basic laptop.  See if you hear a difference...

 

While opinions are nice, there are some actual facts here-optical connections work so well because they are not subject to electrical interference the way electrical wiring is, this just a fact.  It is why optical is the choice for long distance, high rate data transmission.

 

I do not stream from Internet based music sources (Tidal, Qobuz, etc), but for those who do: do you suspect the server(s) these companies are using is an audiophile approved one?  Would Tidal sound better is their server(s) were Innuous?  This is, perhaps, a topic for another thread, but I am just offering it up as food for thought.

 

I tried a pair of FMCs and whilst they made a difference they were never perfect because the FMC closest to the DAC introduces almost as many problems as the optical run solves.

 

I tried this with your MicroRendu and without.

 

The Innuos MKII directly to my Directstream DAC was better on both Ethernet or USB than the MicroRendu was. Fact.

 

The Zenith SE better still. Fact.

 

The Op wants to know whether to buy a MK3, one box solution that meets all his needs, or a server & optical rendu. I'd buy the Innuos in his shoes. But I admit its possible, if unproven so far, that your optical rendu outperforms it.

 

Don't get me wrong, I hope for the sake of the market that the Optical Rendu does perform wonders. That would be a truly great thing.But I'm also aware that you spent ages advocating that servers didn't matter because of the Ethernet isolation into the Microrendu. My own experience begs to differ.

 

It's also kinda disingenuous to suggest you were opining on optical in general when the Op is choosing between your optical product and a specific alternative.

 

Beyond that a topic for a different thread as you say.

 

Cheers,

Alan

 

[barrows]

13 minutes ago, BigAlMc said:

I tried a pair of FMCs and whilst they made a difference they were never perfect because the FMC closest to the DAC introduces almost as many problems as the optical run solves.

Agreed that in this scenario the downstream FMC is the "problem".  But that does not mean that one can still not test the thesis this way:

 

A pair of FMCs and optical cable, then trade back and forth the server for an Innuous to an ordinary laptop.  this test would still show that optical isolation renders the "server" a non issue.

 

The solution to the problem of the downstream FMC, if/when one decides to go this route, is of course to use either a DAC such as the Lumin X-1 with an optical input, or a good purpose built Renderer (like Sonore opticalRendu) with an optical input, as these products should be built to audiophile standards in terms of noise and signal integrity (unlike the problematic FMC in the example).

 

As to a direct (USB) connected server vs. a network based solution that it is a different discussion than what I have been referring to.  Personally i left behind direct connected servers long ago, as there are just way too many problems to address with servers which are much more easily solved by going the Ethernet route.  While I do believe it possible for a direct connected server to perform very well, the extremes in terms of optimization that have to be gone to cost way too much to be practical, and Ethernet audio distribution solves these problems in much more efficient manner.

[matthias]

1 hour ago, barrows said:

Perhaps wrong in your view, but did you try it?

The entire reason to go to an optical Network solution is that it provides complete isolation from the upstream gear.  The "quality" of the upstream signal does not matter (unless it is so bad that you have lost bits, which does not happen on Ethernet transmissions).  An optical connection does not transmit ANY noise from the upstream gear to the endpoint, it cannot.

What does matter is the quality of the signal as it is presented to the DAC, and in this case, yes, the source does matter, but the source is the Renderer, not the computer on the other side of the optical isolation.

This is the entire reason to go with an optical Ethernet based solution.  You can run any type of computer to serve the files, and it will not matter one bit.  Just make sure that upstream computer gear is well isolated on the AC side from the audio system (plug it into a different circuit from the audio system, preferably on the opposite phase).

I really would hope that you are right, I buy a MBP, set it up with an opticalRendu and get the same SQ as with an Innuos Statement. But I bet that the opticalRendu will show exactly differences between upstream servers, their OS, their software players, their settings, their power supplies, maybe even more than the ultraRendu.

We will see or better hear.

 

Matt

[BigAlMc]

6 minutes ago, barrows said:

 

A pair of FMCs and optical cable, then trade back and forth the server for an Innuous to an ordinary laptop.  this test would still show that optical isolation renders the "server" a non issue.

 

No. This is where we fundamentally disagree. The server matters and all this 'perfect isolation' malarkey is, at best, hyperbolae.

 

If isolation, be it optical, Ethernet or captain kirks transponder solved everything then why would you have so many experienced audiophiles spending so much time, money and energy on the Zeniths, audiolinux NUCs and suchlike.

 

Everything matters.

 

Period.

 

And the Optical Rendu might solve say 70 to 90% of it. But it will NOT solve it all.

 

Happy to be proven wrong. But I don't see a queue of Zenith owners lining up to trade their Zeniths in for an Optical Rendu. I do however see a lot of Microrendu owners doing do.

 

Cheers,

Alan

[McNulty]

Many thanks for all your responses. Positively surprised by all the feedback. I'll try to go into some of your suggestions from oldest to newest post.

 

I've checked out the Darko video last week. It appears John used the Innuos as an endpoint only, so not fully similar to my intended use. A bit off topic but I found it odd that around the 18th minute he talks about using the Innuos as a one box Roon solution only in relation to the Squeezelite experimental mode and doesn't discuss the Roon Core mode.

 

I won't use the ripping functionality of the Innuos so indeed this is not a decisive feature. Have also considered Antipodes EX but that model seems a bit overpriced in Europe TBH. 

 

My current streamer is a modded Sonos ZP90 without Roon (so no music server necessary, I do have a Synology NAS for file storage) and my DAC is a Luxman DA-06. Based on the Stereophile measurements by John Atkinson of the DA-06 it seems USB is the preferred input for this DAC (I haven't done a comparison myself). As both the Zenith and the OpticalRendu have USB outputs, this seems to fit nicely.

 

I do have a 2014 MacMini with 8 GB in my home office that could function as a Roon server with the OpticalRendu, but see this as a temporary solution, as I would rather have a dedicated Roon server that I can always leave on. The Audiostore server mentioned by BigAlMc indeed can be an alternative for a SonicTransporter, although they too seem a bit pricey.

 

With regard to the noise of the last optical receiver, I think I kinda got that covered in my current setup by using the optical input of the Cisco 2960-8TC switch before the streamer instead of an FMC with a SMPS. 

 

To sum things up I still haven't made my final decision yet, but of course the OpticalRendu is not released yet and I still have a demo of the Zenith planned, so there's some time to think things over. Again, thanks for all your input so far.

 

 

[matthias]

5 minutes ago, BigAlMc said:

No. This is where we fundamentally disagree. The server matters and all this 'perfect isolation' malarkey is, at best, hyperbolae.

 

Exactly.

 

Matt

[matthias]

1 hour ago, barrows said:

Try any audiophile server on the other end, and then try a basic laptop.  See if you hear a difference...

 

Do you then get the SQ of the audiophile server or the SQ of the laptop?

 

Matt

[barrows]

38 minutes ago, matthias said:

I really would hope that you are right, I buy a MBP, set it up with an opticalRendu and get the same SQ as with an Innuos Statement. But I bet that the opticalRendu will show exactly differences between upstream servers, their OS, their software players, their settings, their power supplies, maybe even more than the ultraRendu.

We will see or better hear.

When you go with this system what goes on inside the opticalRendu is what matters for sound quality.  Essentially the opticalRendu becomes the "server" if you will.  Now, what software you run the system at will change the sound, because different software changes how the opticalRendu works internally: for example: if you run ROON, it uses the RAAT protocol and the opticalRendu will be running as a RAAT endpoint, and so this will sound (a bit) different from if you run MiniMserver which uses DLNA protocol, or if you use HQPlayer and run the opticalRendu as a NAA device-so these Network transport differences for different player softwares will sound different, because the opticalRendu is running differently for each one.

Additionally, if you are doing things like oversampling in the server, or using any other DSP functions, then of course there are differences.

But hardware differences on the upstream side of the optical connection will not matter (unless something is terribly, terribly wrong and you are dropping bits or something, but this is very, very unlikely).  Just make sure the upstream gear is well away physically from the audio system (such that airborne interference will not matter) and plugged into a different circuit, such that the noise it imparts on the AC line is not a significant issue (assuming you have some kind of decent power conditioning in place for the audio system). 

[barrows]

18 minutes ago, BigAlMc said:

No. This is where we fundamentally disagree. The server matters and all this 'perfect isolation' malarkey is, at best, hyperbolae.

 

If isolation, be it optical, Ethernet or captain kirks transponder solved everything then why would you have so many experienced audiophiles spending so much time, money and energy on the Zeniths, audiolinux NUCs and suchlike.

I challenge you to propose a  technical theory by which an upstream device will matter in an optical Ethernet based set up.  Ethernet (wired) isolation is good, but not perfect, and it is true that some interference can get through the small transformers which wired ethernet uses.  OTOH, optical Ethernet is perfectly isolated by definition, electrical noise cannot get through.  As long as we are talking about a bit perfect transmission here (not using different DSP in the server, etc) the upstream hardware will not matter.

 

As to your second question, yes, I agree.  There are a lot of audiophiles wasting time and money on things which do not matter.  This is normal audiophile behavior, and is not going to end anytime soon.  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. 

 

Now, if one is not going to use a Network connected system, for whatever reason, then for sure I would recommend the best possible audiophile type server be used, as with a direct connected server the quality of its output and its noise footprint is going to have a large effect on the sonic performance.

[barrows]

17 minutes ago, matthias said:

 

Do you then get the SQ of the audiophile server or the SQ of the laptop?

 

Matt

neither, it does not matter.  You get the sound quality of the Renderer.

[matthias]

32 minutes ago, barrows said:

Now, what software you run the system at will change the sound, because different software changes how the opticalRendu works internally: for example: if you run ROON, it uses the RAAT protocol and the opticalRendu will be running as a RAAT endpoint, and so this will sound (a bit) different from if you run MiniMserver which uses DLNA protocol, or if you use HQPlayer and run the opticalRendu as a NAA device-so these Network transport differences for different player softwares will sound different, because the opticalRendu is running differently for each one.

Additionally, if you are doing things like oversampling in the server, or using any other DSP functions, then of course there are differences.

But hardware differences on the upstream side of the optical connection will not matter (unless something is terribly, terribly wrong and you are dropping bits or something, but this is very, very unlikely). 

You are walking on very thin ice. Software upstream matters, but hardware not? Are you serious? Maybe you know better than me that you can not separate them in their influence on music. An how can a device play the shades of music when it can not show differences in hardware which have their impact on music?

 

Matt