I’m hearing about folks with Rossini’s using music servers such as the Aurender N20 and using the Rossini only for the DAC portion. They say it improves the SQ notably improved bass and air around the instruments.
Any Rossini, Bartok or Vivaldi owners out there doing this?
What would be your use cases for a device like the N20? What would you like to achieve?
I had the N20 at home for two weeks some time ago. Music data was delivered to a Bartók through the S/PDIF interface and the Vivaldi clock was connected to both the Bartók and the N20. N20 has no ethernet output.
Listening dispositions and hifi aims are different from individual to individual. Since it’s a hobby I believe there’s no right or wrong. Just suitable or not suitable. I felt the N20 made the music more audiophile. (I use the term in a neutral way.) When I unplugged the N20 at the end of the two weeks there was a half hour moment where I readjusted to just the Bartók with clock and after that was happy.
Audiophile sound is a subordinate aim for me. It’s more about how deep can I listen into the music, feel the intent, go beyond cognition whenever possible or just rest in music. An external clock helps achieve these requirements.
Audiophile sound can have a sugar coat effect on certain music. It usually enhances benevolent music. That’s a good thing. When music gets abrasive, aggressive, sinister, stormy however I want raw and scary. Not a growling front man with a fluffy silencer between him and his mic.
By now I settled on an Innuos Statement Server for offline music. It is connected via Ethernet to the Vivaldi Upsampler. The Upsampler itself serves as streamer and feeds into Vivaldi DAC.
Others here use e.g. Melco, Roon Nucleus, Intel NUCs for music servers. The choice first depends on the use cases and secondly on ones listening dispositions.
Marco,
Thanks for your thoughts. I guess I wasn’t aware that a music server could make a noticeable difference, whether good or bad, in SQ. I can understand a difference in SQ a different DAC may make. Does the Roosini not do a good job in presenting the data bits to its DAC?
Thx,
Brian …
Hello Brian,
the terms server, streamer and DAC are being used loosely in the hifi field. Maybe I attempt an overview. Hope it’s not too dry…
A music server retrieves music files from a storage device. The server can be located in ones home (e.g. Roon Nucleus, MinimServer running on an Intel NUC) or it can sit in a data center somewhere in the wider vicinity (e.g. a mirror of the Qobuz servers). The files might be stored inside the server or in attached storage. The retrieved music files represent the original analog signal. The server takes the music file, slices it into small pieces and wraps those pieces into a transfer protocol. These protocols at home could be RAAT for Roon or UPnP for standardised media data transfer. In online streaming there are proprietary protocols of Qobuz or Tidal and so on. The protocol is necessary for transfer as the whole file can not be sent in one fell swoop. Also, networks based on the Internet Protocol behave more like London city traffic than two tin cans connected with a wire. Alas traffic management is necessary, which the transfer protocols help with.
A streamer or player (digital to digital converter) receives the data stream coming from the server and processes it further. The transfer protocol is unwrapped, the original data reconstructed and then readied for the DAC. This is because the DAC in itself can not process music files the way they are encoded for storage. A stored file might have a FLAC or WAV format but the DAC needs an input of a PCM or DSD stream. At this point the output needs to be synchronous too. Meaning that what gets sent is delivered without delay. Because of this, now jitter becomes an issue.
The DAC (digital to analog converter) in the end receives a time sensitive stream of PCM or DSD data and then turns it into an analog signal that feeds into the amplifier. There are some marvelous articles on this by dCS staff on this forum. For example here. https://dcs.community/t/dcs-ring-dac-a-technical-explanation/2724
The whole thing is a bit confusing, as the three processing stages of server, streamer/player, and DAC are being put into single or multiple boxes by different manufacturers. dCS manufactures no server. In Bartók and Rossini the streamer and DAC are in one box. In Vivaldi the streamer is inside the Upsampler device separate from the Vivaldi DAC.
If you use an Aurender N20 you can store music inside it and you will always bypass the dCS streamer and go directly into the DAC. The N20 is a server + streamer/player. In my ears the dCS streaming components are more than capable. I see no need to bypass the streamer. A server becomes necessary as soon as you have music stored on drives at home.
There’s somewhat of a consensus among forum contributors that Ethernet is the best way to feed the dCS components. If you use Roon, you need a component to run Roon core on to perform the server duties. That Roon server then handles offline (locally stored) and online (strored in a data center) music streaming. It feeds data wrapped into the RAAT protocol into the dCS streamer, which is able to unpack the protocol (Roon ready). If you rely on Mosaic app on the other hand, online streaming like Qobuz works out of the box. Offline music streaming then needs a music server at home which runs a UPnP server softfware on it. That’s where the Melco or Innuos or Lumins come into play.
Very long story short - I have tweaked the system delivering online music streams so that I am happy listening to the streams and don’t miss offline music. Except for the important factor of genuinely owning a music collection… Tweaking entails using good external power supplies for network gear etc. Offline I tried an up to date Synology NAS (Network Attached Storage) device running the MinimServer UPnP software on the NAS itself. In that configuration the Synology NAS became the music server. I was dissatisfied with the result. I kept switching back to online streaming in that setup. Now with the Innuos server all is fine. Maybe this all was a long winded way of saying I have no good explanation why offline servers sound different. In my ears they do. It’s not day and night using the Vivaldi Upsampler as the destination. But the difference was enough for me to switch one off pretty soon and enjoy the other thoroughly. Now my offline collection is the slightly preferred place over online streaming to source an album from. Maybe with the fancier server I can listen deeper into the fabric of the music, where with the NAS music was more closed for inquiry.
Kudos to Marco, but essential to understand is:
On the face of it, playing and streaming digital music files is a straightforward process. You direct data from various sources—some local, some “in the cloud”—perhaps via a reclocker/signal conditioner to a digital-to-analog converter (DAC). “And the music comes out here.”
Not so simple. Bits, it seems, aren’t bits, or not only. A digital datastream is also an analog signal. Noise and other signal errors endemic to multi-function computers not designed primarily for music playback can affect how music sounds. And then there are the practical issues of setting up and connecting everything optimally, and then organizing music files correctly, which can be especially difficult when ripping files from multidisc sets.
These obstacles are why so many audiophiles have either switched to one-box music server solutions or even thrown up their hands and stuck with physical media.
From: Stereophile | Media Server Reviews | Sep 29, 2021
Marco,
Thanks for a thorough explanation. Most of my source is Qobuz but I do have a Synology NAS for my offline tracks ripped over the years (mostly FLAC).
Are you saying that for my offline tracks, I may want to audition an Innuos music server? If yes, which Innuos model do you suggest?
Thanks,
Brian…
Hello Erno, thank you, yes, shorter is better. 
I try a shorter wrap up now. Streaming from server to streamer via internet protocol suite should be immune to most things by design. After all, the data is delivered bit perfect and layers of abstraction make the delivery clocking insensitive. And still - a fraction of people myself included hear differences between switches and music servers. So much that they spend hard earned money.
Probably it comes down to this - we connect systems to systems and the system boundaries are somewhat permeable.
Hello Brian,
yes, if the device is easy enough to obtain, I would go for it and try at home. The interesting questions when listening could be “How are Innuos and Synology presenting music in comparison?” “After listening for a while, which source do I prefer?” It will be probably less about traditional hifi criteria. More the enjoyment factor.
I heard the Innuos Zen Mini at the dealer and found it to be a competent source in his system. Then I tested the ‘Statement’ at home and enjoyed it enough to buy it. The other models (Zen, Zenith) I have no experience with.
Assuming that their lineup is ordered by sound quality I would opt for the device that fits into your overall system investment. Maybe you can listen to more than one model at the dealer and listen to your preferred model at home afterwards…
A very interesting thread. I have several friends who are using Innuos servers with their dCS and Naim systems. If possible I would like to expand the conversation and speak exclusively about online streaming (Qobuz and Tidal) vs. local playback of ripped files. Up until now I have held back from an external solution such as the Innuos since the overwhelming majority of my playback is via streaming services (despite having 3k CD’s - many of which I have yet to rip during my past 10 years of digital playback).
Prior to acquiring my Rossini last year I used a Linn Klimax DS (which now resides in my headphone systems) and Linn, like dCS, has always advocated the use of ethernet and a bog standard NetGear ethernet switch. Things have now changed with the introduction of the new Organik DS and the inclusion of an optical ethernet input. Linn go as far to say optical is now preferable and will result in enhanced performance vs. standard copper ethernet.
In the context of online streaming only what optimizations can be done to the Rossini - or are we still to follow dCS recommendation of an unshielded CAT6 cable with a compliant network switch (in my case a Cisco 2960).
Best
Gregg
As far as dCS recommendations are concerned it is to use unshielded ethernet cable that is certified as compliant with a CAT protocol. So currently, given what is available on the market , this would mean certified CAT 5e or certified unshielded CAT 6.
As far as streamers and switches are concerned it seems that from an engineering viewpoint there is no ready explanation of how or why an audiophile NAS or switch would be preferable to a standard IT one. However all of the respondents to this forum who have tried switches report a change in the resulting sound. Of these one felt it was impaired ( using a Melco S100) whereas the majority have reported noticeable improvements ( most but not all using Melco S100 as well).
So there are currently three areas of interest none of which " improve" Rossini itself (that is left to the addition of the matching clock, and perhaps power cables/conditioners /fuses etc.). Others are held to improve the holistic presentation of the data to the DAC in some way ( i.e everything being accounted for not just accordance with a conceptual model of data transmission). One area has an engineering basis ( the use of unshielded cable) but two of which come with no provable explanation at present despite reported subjective positive outcomes ; servers or switches.
I could only suggest that you may like to experiment with the latter if you can arrange access to them. If, however, you are of the mindset that requires a ready scientific explanation first then you might be wasting your time.
Guys,
I stream from Tidal and from my Synology NAS (thank you PAR) using Mosaic on a network bridge. Ethernet cables all same. All go thru 3 switches (SOtM + 2 EtherRegen).
No diff in SQ that I can hear once the NAS power was upgraded. It has a 4 amp requirement, so be careful when searching for a PS, if that is your intent.
By the way, I have found that iFi smps’s work better for me than any of the LPS’s I have used. The level of convenience and price is another level altogether when compared to the LPS.
Best wishes
Aubrey
Marco, I have 2 questions.
Is the Innuous Statement similar to the feature/function that an Aurender N20 provides? I ask because I have an opportunity to purchase an N20 at an attractive price.
Would I be better off purchasing a Rossini Clock instead of a music server for my Rossini?
BTW, I recently upgraded the Rossini power cable to a Shunyata Omega QR-s cable. It’s amazing what a high end power cable can do.
Hello Brian,
to answer your first question: the Innuos Statement is similar to the N20 in features and there are some differences. Both are servers and streamers. With the N20 you have to use the streamer, with the Innuos you can choose not to use the streamer with the ethernet output. In that case the Innuos is either an expensive well made NAS with a UPnP server software running on it or it behaves as a Roon Core. Things the Innnuos has, the N20 doesn’t: CD drive to rip music, UPnP server to deliver music via network to compatible destination devices be it e.g. dCS or Sonos or Linn, run Roon core to enable using Roon with the Rossini. And now the N20 ‘haves’: SPDIF and AES/EBU digital out, wordclock input. The clocking I tried with the N20 and it is cumbersome to use. The N20 only has one clock input. So you have to change sample rates at the clock if necessary.
Your second question is harder for me to answer - server or clock first. I don’t know how you perceive changes in the digital chain that sits between Rossini and the source file - irregardless where the source file is, at home or in a data center. Some people hear differences, others not. Both is fine.
You could do an experiment. Choose albums you like and know well. Compare how they are presented to you when coming from your NAS and when coming from Qobuz using Mosaic. They might sometimes be different source files, but if you listen to enough of the music you get a feeling. Don’t a/b test, listen to whole albums. The question then is ‘Does a preference arise?’ If you enjoy Qobuz more than music played from the NAS, buying a server is a good idea. There are lots of ‘moving parts’ in this comparison though. What router you use, which switch, which power supplies, and so on.
In my setup now, I like streaming and offline mostly the same. The sound signature is different. Before that I chose Qobuz over Synology every single time. It’s good to observe how we use our system in the end. If we gravitate to online streaming and neglect the same offline content, it might tell us we enjoy online streaming more.
Maybe - as I think while writing - your upgrade path could be clock, switch, server or switch, clock, server. The dCS streaming board is very competent in delivering music to the DAC. I don’t see a necessity to use a different streamer. Source wise then it’s down to the server. The clock improves the Rossini’s ability in total. Both Qobuz and offline server will benefit. The music will be more ‘there’ and you more ‘in it’.
Good you upgraded your power cord. They can make a huge difference. One needs to swallow the fact though that cables are priced value based, not cost based. Here’s a nice article on the principle:
Hello Gregg,
the optimization upstream from the Rossini is a rabbit hole suited for computer nerds. General advice is difficult as our respective network setups are so very different. There are a myriad of solutions possible.
What made a difference for me?
Using a switch made by a hifi brand.
There’s a lot one can do around power ordered by decreasing impact on sound quality: power supply, power conditioner, power cable, power distributor. All those work in both directions - better power input into the device, less contamination out of the device (modem, router, firewall, switch, etc). As the first device to get a power supply for I recommend the switch. When I added it to my switch it was a wow moment - thaaat’s what the power supply can do?
And now I might be stoned to death. Resonance control of network gear works for me. The effect is a lot smaller than upgrading power, but it adds to the naturalness of sound. If that’s too far fetched for some readers kindly forget I said it. 
The network cable connecting my Vivald Upsampler to the wall socket is a compliant unshielded non audiophile cable as per dCS recommendation. It actually sounds better than the very expensive cable I had before. In the basement I still use hifi network cables. Their effect is smaller than power, but here we need to think about the whole as Pete alluded to. They help make the sound more analogue.
Switch, power, resonance control, cabling are the four things I improved. If you get the whole optimization right, the prior status quo sounds as if something is broken. Seriously.
I have nothing optical yet, that’s the next step when time and inclination permit.
If you’re interested in the gear choices…
As stated above, it’s a
I recently took the plunge to digital, ripped my collection to AIFF, and sold all of my CDs. (I’m in NYC, and therefore have space concerns, and yes, it was painful!)
I have been wonderfully impressed by this signal path:
All music stored on 4TB SSD inside Roon Nucleus
Roon Nucleus → Ethernet
Ethernet → DAC
DAC → Power Amp
Power Amp → Speakers
I like simplicity, but I realize I am in the minority on that one…
Marco,
I can’t thank you enough for the insightful suggestions and the article on ‘Value-based Pricing’.
Re: my Ethernet equipment setup, I use a Ubiquity UniFi Switch 8 60 watt switch (https://store.ui.com/collections/unifi-network-switching/products/unifi-switch-8-60w?gclid=CjwKCAjw_L6LBhBbEiwA4c46uj00s3V2_A6Qep0EhJXineeC6HkBGEenEpkPGoyqZid0yNIh43E_XhoC1qgQAvD_BwE) and a pair of TrendNet Fiber Media Converters to try to eliminate any electrical noise I may have coming from the 75’ CAT6 cable from on our upstairs office to my downstairs listening room. The second TrendNet Media Converter then connects to the Rossini via a certified BlueJeans CAT6 Unshielded cable. The Ubiquity switch and the pair of TrendNet Media Converters and all audio components are grounded to my Shunyata Everest power distribution box. The power supplies I use for the Ubiquity switch and TrendNet media converters are just the stock wall warts.
Your suggestion to experiment listening to the same albums via Qobuz and locally ripped albums makes a lot of sense. Just so I’m clear, are you saying that if I enjoy music using Qobuz more than my Synology NAS, then a music server may help. I assume the music server replaces the Synology NAS, correct? Does this also mean that a music server has no impact on the SQ of Qobuz played via Mosaic?
Thanks,
Brian …
Marco,
I can’t thank you enough for the insightful suggestions and the article on ‘Value-based Pricing’.
Re: my Ethernet equipment setup, I use a Ubiquity UniFi Switch 8 60 watt switch (https://store.ui.com/collections/unifi-network-switching/products/unifi-switch-8-60w?gclid=CjwKCAjw_L6LBhBbEiwA4c46uj00s3V2_A6Qep0EhJXineeC6HkBGEenEpkPGoyqZid0yNIh43E_XhoC1qgQAvD_BwE) and a pair of TrendNet Fiber Media Converters to try to eliminate any electrical noise I may have coming from the 75’ CAT6 cable from on our upstairs office to my downstairs listening room. The second TrendNet Media Converter then connects to the Rossini via a certified BlueJeans CAT6 Unshielded cable. The Ubiquity switch and the pair of TrendNet Media Converters and all audio components are grounded to my Shunyata Everest power distribution box. The power supplies I use for the Ubiquity switch and TrendNet media converters are just the stock wall warts.
Your suggestion to experiment listening to the same albums via Qobuz and locally ripped albums makes a lot of sense. Just so I’m clear, are you saying that if I enjoy music using Qobuz more than my Synology NAS, then a music server may help. I assume the music server replaces the Synology NAS, correct? Does this also mean that a music server has no impact on the SQ of Qobuz and Mosaic?
@keiserrg You raise a topic that most of us have or will go through, the decision whether to go direct or use a pre-amp. For a long time I used my Rossini as a pre-amp and went direct to whichever amp I had at the time. There are some interesting benefits of going direct such as a slight improvement in details and less colored music. But I found I was less engaged as the music seemed a bit too sterile. I tried some high end pre-amps such as as AR Ref 5SE and even an AR Ref 10 and still preferred the direct from Rossini to amp option.
At the suggestion of a trusted friend I decided to give the pre-amp option another shot. This time I also took my friend’s advice to properly mate a pre-amp with the amp. My current amp is a Gryphon Antileon Evo Stereo so the obvious choice for a matching pre-amp is the Gryphon Pandora.
I couldn’t be happier with the results. I’ve never experienced sound from my system so life-like and engaging. With the pre-amp in place, I now look forward to listening sessions a lot more. The Pandora has another benefit in having a world-class phono section. I never knew my Project RM-10 turntable with an average cost Sumiko Blue Point Evo could sound so good.
Those are some serious electronics Brian, congratulations! I’ve heard great things about the Gryphon line but have never listened to it. What speakers are you driving?
I am wondering, did you ever try the Rossini direct to a tube amp, and if so, did some of the harshness you mention fade?
The speakers are Rockport Altair 2’s. I never tried a tube amp with the Rossini. The sound I experienced when going direct from the Rossini to an amp wasn’t harsh but it was just a bit lifeless. Some describe it as not having enough ‘meat on the bones’.