What makes One Music Server Sound Better than Another?


So this week my Mojo Audio DejaVu music server that I have used for the past 2-3 years crapped out. Benjamin at Mojo was more than helpful and the DejaVu is on its way to Mojo Audio where it will make a full recovery.

Thankfully, I still have my Antipodes DX2 Gen 3 (their former flagship) music server so I hooked it up. After wrestling with Roon protocols, transfers, and set-up menus, I was able to get it going so I have music. The DX and my Sonore Sig Rendu SE opt. are both connected to my network so the DX (like the DejaVu), is only being used as a Roon core and the Sig Rendu SE serves as the Roon endpoint for streaming Tidal and Qobuz, with a direct USB connection to my DAC.

The point of this thread is to ask, how come I perceive the the DejaVu server as sounding better than the Antipdes DX? In fairness, the differences I perceive are not great but it seems the DejaVu is fuller sounding, more tonally rich, and bolder. Is this why some here spend $10K+ on a Grimm, Taiko or something else?

If a server is basically a computer, sending digital information to a streamer/endpoint and, assuming that digital information is transmitted asynchronously and reclocked by the DAC’s master clock, and assuming noise is not the issue (i.e., both units are quiet and there is an optical break between the network and both the server and endpoint) then what are the technical reasons one should sound better than the other? It is not that I want to spend $10K+ on a music server with a lifespan of maybe 5 years before becoming obsolete, but I would like to understand what more you are getting for your money. So far, the best I can come up with is lower internal noise as the major factor.

As a side note to the above, when I thought things looked hopeless for getting set up, I scheduled a support session with Antipodes and, although I lucked into the solution before the meeting time, Mark Cole responded ready to help. Setting up the session was super easy and reminded me of the superior level of support I had come to enjoy from Antipodes during the time that the DX was my primary server, including multiple updates and 2 or 3 hardware upgrades, which prolonged the service life of the DX. Good products and good company.

 

mitch2

Showing 14 responses by mitch2

@mapman 

  • Both servers use a version of Linux, maybe not the same - I wouldn't know
  • In both set-ups, the server (DejaVu or Antipodes DX) is what Roon considers a Music Server running what is called Roon Server in the Sonicorbiter app (essentially, it is serving as the Roon core, only)
  • In both set-ups, the server has only a single direct connection with the network/router, and no connection to the streamer or to the DAC
  • In both set-ups, the Sonore Signature Rendu SE optical is what Roon considers a Network Player (or streamer) and is running what is called Roon Ready in the Sonicorbiter app 
  • The Sonore Signature Rendu SE optical is NOT connected directly to the server but has a single direct connection with the network/router
  • The Sig. Rendu SE opt. is connected with Mojo Audio's Mystique EVO Pro DAC via a direct USB connection currently using Network Acoustic's muon USB cable

 

@rbstehno 

"I have my server in another room and it sounds superior to using usb."

Exactly.  The server, modem, router, etc. are in my network room and the streamer and DAC are in a separate room with my audio system.  The streamer connects to the router by 45 feet of fiber and USB connects the streamer to the DAC. 

@tonywinga 

Your K50 is a server/player, performing both server and player functions before directly feeding your Aries-Cerat Helene DAC.  Since the player is in the same box as the server, and since the K50 server/player resides in your equipment room, it needs to be quieter than a server only component that resides at a location away from your analog gear.  I have no doubt it sounds excellent.

The interesting experiment would be to first have the K50 in your system room connect directly to your network, perform the player-only function, and output directly into the DAC.  Second, set up a second K50 in a separate room, also connected to your network, and performing the server-only function.  Finally, replace the server-only K50 with a different, lower priced computer-type (i.e., not audiophile royalty) server, although still using the K50 in your system room as the player, and then listen for a difference in sound quality.

 

Tony, the point was not really about how it sounds in your system (I suspect it sounds fantastic) since both the K50 and the Aries-Cerat Helene are well-regarded, top notch digital components. 

The question is whether the extensive build properties of the K50 (and resulting price) would be necessary if it were used as a server-only, or whether the same sound quality could be achieved by using a (less expensive computer-type) server located in another room away from the system, performing server-only duties, and feeding a K50 used as a player-only into the Aries-Cerat Helene DAC, i.e., would you be able to tell the difference between the server-only K50 and a different server?

@lalitk - Good to know about the longevity of the Aurender. Only downside for me is that I don’t plan to give up Roon so I need something that works easily with Roon. Still thinking about the K1, since I only need it to perform the server function.

To your implementation comment, when I opened it to take pictures, I noticed a lot of shielding and compartmentalizing inside of the DejaVu server. Benjamin at Mojo is big on the small details.

One thing I still don’t quite get is the folks associated with audio high end, such as Atkinson at Stereophile, who use a Roon nucleus server. That type of observation brings me back to my question, if those sound just as good at less than $3K, then why do others spend north of $10K?

@ricred1 - No need to be sorry, the praise is deserved IMO.  The K41 would be my sweet spot based on how I am set up but I have watched these things over the past 10 years and the service life of these servers seems to be south of 5 years before they become obsolete, making it an expensive proposition.  Antipodes did better than most at prolonging the life of my DX by offering the 2 or 3 upgrades I had them perform, and of course - it is still working as I sit here and listen!  Manufacturers like to say they have created a modular design for easy upgrades but regardless of how they initially sell it, you most often see an entirely new product are rarely a true upgrade opportunity.  Antipodes did a lot better than most.

I will also give a shout out to Benjamin at Mojo Audio, who also responded very quickly and based on pictures was able to let me know the DejaVu can be repaired.  He is another who will upgrade existing units until it doesn't make sense.  Unfortunately, he is no longer making servers, but I sure like the musical sound of his DACs.

@yage - You got me...of course it was sighted since I had to move connect the DX after the DejaVu failed.  Yes, there is the possibility of no real SQ difference except what is in my head. 

However, I am still interested in what others say because if the no difference path is true, then why are some paying so much money for their servers.  This is NOT a thread to bash those who do pay a lot for their server but rather to help me understand why.  I am interested in hearing the technical reasons why there should be differences in the case of a server only assuming the unit is reasonably current with an adequate processor, appropriate software, reasonably quiet power supplies, and SSDs if internal drives are used.  I am curious what I would be getting for my money if I were to spend $10K or more.

Thanks all.

It is an interesting comparison given that both servers were positioned and connected in the exact same way to the exact same equipment. Isolation is provided by an optical breaks between the router and server, and also from the router to the Sig Rendu SE opt. streamer, and LPSs are used on all the converters and network equipment. Therefore, the incoming signal should be completely isolated from the audio equipment by optical fiber. Based on the discussion, I will conclude what I hear may result from:

  • Improvements in processing between the older DX and the DejaVu, coupled with increased demands in running Roon,
  • Improvements in isolation of components within the DejaVu over the DX, and maybe in the implementation of the network connection,
  • Finally, maybe I have a bias in what I believed I was hearing. The more I listen to the DX, the more I am ok with what I am hearing.

In any event, the DX is older but would still work quite well for somebody. Mark Cole offered to update the processing remotely and maybe I should take him up on that and then listen for differences. I could then more easily sell it when I get the DejaVu back or move to something else.

The other thing I seem to be learning is that this digital thing is sort of a crapshoot. Server, streamer, DAC, all play a role but it seems to be hard to pinpoint the impact of each on the overall sound. Adding the Sig. Rendu SE opt. seemed to elevate the SQ I heard in specific ways. Also, changing out DACs results in a clearly and reliably audible SQ difference in my system, which surprises me less because of the DAC's job of converting from a digital to an analog signal. I probably shouldn’t be surprised that different streamers sound different from each other, but I still don’t understand why.

@audphile1 - Yes, the two servers I am comparing were at approximately the same level, when new, although the DX is older technology. The DejaVu seems quite well built with regard to spacing and isolation of internal components, isolation of connections, and power supply. All of these things may be adding up to the fuller, richer, sound I perceive through the DejaVu, maybe?

I am sure the WiiM Pro would be an interesting comparison but it is a player/streamer and the comparison I am making is between the two servers/Roon cores. I already noticed a sonic difference moving from my Metrum Ambre player/streamer to the Sig. Rendu SE opt., so I am certainly convinced on that front.

@mapman - No DSP here, just 24-bit, 192kHz, PCM into an R2R DAC, or 24-bit, 96kHz, PCM into the older technology SMc DAC-2 (delta sigma). I have a Benchmark DAC here and could try DSP, but I really have had no desire to try that yet.  Do I understand correctly that DSP can put additional demands on the server processor?

@audiotroy

just isolating the clock board enables the server to produce a larger soundstage, greater clarity, and a more analog like presentation

Still curious how something like isolating the clock board affects those analog sonic traits when sending a digital signal. The digital signal is still read at the other end, corrected as needed, reclocked, and converted to analog. Does isolating the clock board somehow modify the digital signal that gets sent? Does it affect jitter? What is the technical mechanism for these changes? Not trying to be difficult, and I am certainly not saying it doesn’t sound different, but I would sure like to hear a technical explanation of how that isolation changes a digital signal.

@mapman - Your points are valid and I agree but they mostly do not seem applicable to my specific situation.

  • the server and streamer are independently connected to the network through a wired Ethernet connection, and each of those two wired Ethernet connections have an in-line fiber optic cable between the component and the network connection
  • all network and peripheral components are powered by LPSs that are connected to their own dedicated 20A circuit
  • in the case of the streamer, there is only a very short Ethernet cable at the router end and then an optical converter and fiber optic cable extending the entire distance to my system room and directly into the streamer (i.e., the streamer itself has a fiber optic input)
  • All connections to my main system are wired and do not use wi-fi, however, my auxiliary (living room and outdoor) systems are connected by Roon endpoints directly wired to Orbi mesh satellites, but those systems are not the topic here
  • there are no noise issues that I discern
  • the DAC is a $10K (when new) model with a fairly high quality JLSounds USB input used by other high end manufacturers

The issue is not that my system doesn't sound good, but that it sounds better with one server than with the other and I am curious why that is.

@kennyc

"Is the OP asking why better parts and/or superior implementation sounds better?"

Not exactly. Both the servers I recently compared were designed and constructed by manufacturers known for fastidious attention to detail in the areas of power supply, shielding, and isolation, and both were the flagship unit for those manufacturers at the time they were built. Both have "better parts and/or superior implementation". One is a bit older (2014 era.) while the other is a little under half that age. The installation of both servers in my system was exactly the same and included fiberoptic isolation between server and router, and between endpoint and router, which should mitigate electrical noise on the Ethernet lines.

I perceived a bit fuller, tonally richer, and bolder sound from the newer server (which is headed back to the manufacturer for repair), and I am curious what aspects of the design or implementation would result in those sonic attributes being different between the two servers, assuming what I perceive is real and not a bias of some sort. I am trying to understand "why", considering this is a digital signal, one would sound different (and better) from the other, when connected to the exact same receiving equipment?

Based on the discussions I have read here and elsewhere, I am left with the impression that the reasons for the sonic differences are not fully understood but that processing, isolation (mechanical and electrical), and power supply are likely affecting the digital signal (maybe wrt to timing and lower jitter) that is received by the endpoint, and that these effects result in a slightly different analog presentation after the signal is converted in the DAC. In short, it seems the better designers have improved their implementations over the years so I should trust what I hear and purchase what sounds good to me, not unlike most stuff in this hobby.

Thanks all for the responses.

 

 

@jji666 

"It’s about server streaming to endpoint over a network."

That is exactly what I am doing, along with having optical fiber isolation of both the server and the endpoint from the network.

and

"As long as you have bit perfect music input at the server end you’re going to have the same bit perfect output, no matter what else is at the server end, at the streaming end."

This is the presumption I am going by.  My plan is to continue using a reliable, well-built, server, somewhere south of $5K, and then spend a little more as needed on the endpoint and DAC.  I may actually be there now with the Sonore endpoint and Mojo Audio DAC but, in either event, I do not plan to purchase a meg-buck server to chase potential SQ improvement that may not actually be realized for the server-only function..  

@jji666

"My personal view is people buy these pre-built audio server computers because they look cool."

That may be for some but I suspect it has as much to do with people being concerned with upgrading every possible component in their system to the highest possible level so as not to miss out on the n’th degree of sound quality.

My server sits in a back room with my network equipment so from my perspective it could look like a cardboard box, as long as it works well.

To the point by @fastfreight about skepticism related to digital front end equipment, I think that is related to our preconditioning that 1’s and 0’s are relatively immune to outside effects so as long as the 1’s and 0’s arrive at the "receiver" intact, the resulting isolation, error-correction, and processing will handle it from that point.  In contrast, most seem to believe that the downstream analog processing (i.e., D/A conversion, amplification, and loudspeaker conversion of an electrical signal to a sound wave) is not immune to outside effects, and therefore believe the music we hear is largely dependent on the quality of the DAC, amplification, cables, and speakers.

I have no opinion on switches except for my own experience owning two Silent Angel Bonn N8 Ethernet Switches.  These are essentially the same as the English Electric 8Switch as both are made by the company ThunderData and use the identical board, with the cases being the main (only?) difference.  In my system, the Bonn N8 switches work great as switches but they make zero discernable sonic difference.

I am however curious how fiber optic transmission of a digital signal could possibly limit dynamics, and also how a silver plated Ethernet cable could possibly affect transmission of a digital signal to the point of audibility.

@krell_fan1 

"I believe this thread is which SERVER sounds best, as opposed to which STREAMER?" 

You are correct that the original intent was about servers not streamers, but not so much about which server sounds best as about trying to understand why they sound different from each other at all.  

After spending more time with my back-up Antipodes DX server, I am not convinced that I hear much if any sonic difference between it and my Mojo Audio server.   I have however found that I perceive sonic differences between streamers which, if true, would imply that it might be more likely to perceive sonic differences between one-box server/streamers, such as Grimm MU1 or Antipodes K50 than between stand-alone servers.  This would be especially true if the streamer is processing the signal in some way.  My own Mojo Audio is a one-box solution that includes a JCAT USB card but I have found it sounds just as good when used as a networked server-only, being played through separate streamer/endpoints.  I will soon be trying a SGC sonicTransporter i9 optical (Gen 3) so I will then be able to compare the SQ I am getting between three different units used as servers-only; the Mojo DejaVu, Antipodes DX2 Gen3, and the sonicTransporter i9 optical (Gen 3).