Good point. If there is an audible difference between Roon and RoonServer then something is propagating over the Ethernet connection to the PS Audio output.
Here’s how I can break it down in my head theoretically. Feel free to correct.
-
There is simply no difference.
-
The computer and the output device are close to each other. We can solve potential airborne interference issues by moving the devices very far away. Ethernet solves this.
-
There is electrical noise on the ethernet line caused by the iMac (whatever that may be) and it gets passed the pulse transformers in the chain and into the Output Device. If this is the case, then let’s use a fiber media converter.
Sure, one might argue that the FMC’s power supply introduces its own noise, however, the noise should be the same regardless of whether Roon or Roon Server is used, because, we’ve isolated it by fiber. So the source of noise that the Output sees is only the FMC’s power supply.
- Even with the FMC present, there is still difference with Roon and Roon Server. In that case, it would have to be a networking issue where there is an additional load (affecting analog performance) created on the Output device as a result of the digital side being less optimized with regards to Roon vs. Roon Server. So the root of the problem will be a digital one.
That’s the 4 theoretical scenarios I have come up with. I think most will say it’s #1 and move on. However, if we assume Janos is correct in his findings, then #2-4 is what I have at the moment.
Hope that made sense.
I think there is another kind of theory that suggests switching power supply of the noisy computer corrupting the AC power source more due to increased power load and noise, which in turn affects audio equipment. So I think it is not uncommon for people in computer audio community to replace their computer and NAS power supplies, and adopt a fanless design. Disclaimer: I’m not saying this is true or not (topics like this can become very controversial).
Very fair points Peter.
In my case, my Hugo2 is running on it’s batteries during critical listening and the USB input is properly isolated (power, ground and data lines) from everything upstream, with the Uptone Iso Regen + LPS-1.
I’m running RoonServer on the Mac (I just discovered today, thanks to the earlier tip) and Roon Remote on the iPhone - but the kicker is my modem’s WiFi is always off when I’m in the office - the iPhone plugged into the same powered USB hub as the USB-ethernet adapter, and the Mac also connected to this hub. And Roon Remote on the iPhone can access RoonServer via it’s lightning cable - no WiFi. This is producing interesting early results.
I had no idea the iPhone could access RoonServer via it’s lightning cable, with WiFi switched off. The iPhone is nearby (not too close of course) at all times on the desk anyway (for work and family emergencies) so having the iPhone connected to the network via cable presents a new opportunity…
Previously CPU usage with the full-blown Roon app installed was ~15%. Now with only RoonServer installed on the Mac, CPU usage is ~5%… I have no idea about the changes in GPU usage but logic would say it’s down when Roon Remote is not open. And Roon Remote only needs to be launched on the iPhone when changing album or playlist or hitting next track etc. I also have the option of launching Roon Remote on the Mac whenever and then closing it when not needed.
More proper testing this weekend…
All equipment will put some noise back into the mains. So, the ethernet signal route is not the complete isolation step that it might be. An electrically noisy PC in the room next door can be on the same power circuit as the sensitive kit that the user thinks is isolated.
The network router itself can be a noisy device and is often quite close to the playback equipment.
Many things to be considered when chasing down the reasons for observed differences.
1 Like
So I’ve actually done the opposite of what @danny suggested.
Rather than up the CPU usage while listening to A+, to compared with Roon, I’ve downed the Mac’s CPU usage while listening to Roon, as detailed a couple posts up.
Even without a scientific experiment, I am liking it. And I can’t think of a compromise made with this arrangement, with the way I listen. I may have actually scored the rare free lunch…
Lucky because I was about to purchase a faraday cage for my Hugo2 or Macbook… @brian
Okay, sure let’s continue with this thought of train. Let’s add the power thing as well. (I have a Torus isolation transformer for this purpose.)
What’s the difference between Roon vs. Roon Server if you open up Roon and then just press play on an album and then close the Roon window?
I tried it and couldn’t notice a difference if the Roon’s UI was open or not. On my computer, when Roon’s window is up and running with scrolling, the CPU can go between 15-20%. When I close the window, with just the music playing, the CPU is at 4-6%. I didn’t notice a difference between the two.
How would Roon Server differ from Roon with its window closed?
If you have an older OS X on Mac, you may want to try Amarra’s optimization script. I used it 3-4 years back. It would ask you to disable a bunch of things and I remember I would need to restart the computer for it to fully take effect. It actually helped with the sound quite a bit. I was using USB at the time. I remember it as being even more aggressive than Audirvana’s system optimizer.
Found it: http://www.sonicstudio.com/amarra/howtobuildaserver.php
Noted. I guess I created this thread because I heard differences and some others shared a similar experience.
I guess it doesn’t mean there is a difference just because a few of us think there is - that’s where Danny is coming from regarding his ‘keep it to yourself unless you have stats’ vibe.
On the other side of the coin, if there is a difference, it doesn’t mean everyone will hear it - system specific? hearing specific?
It’s more than just closing the window. It’s quitting Roon Remote completely on the Mac, with only RoonServer running in the background. I would assume that there is lower GPU usage when quitting Roon Remote app. See earlier responses from Danny and Brian regarding CPU/GPU usage.
May or may not.
What I mean the ethernet cable impacts SQ. So what economically feasible on a short ethernet cable may not be for a long one. I likely would not be able to afford a 50ft long Purist ethernet cable, but could live with a cost of 6ft long. If I switch to an “ordinary” ethernet cable that itself impacts SQ.
Also, to make a fair comparison would need to get a second PS Audio P3 regenerator just for the Mac mini if I put it far away (I use a PS audio P3 regenerator on the front end)
Thanks, AJ. Interesting that they just don’t send the bits out TOSLINK unaltered, especially at 44.1. The conversion to 48k, though not bad, lacks the space and image resolution of the Plex/Roon conversion process. It is noticeably inferior.
Last night, I played Countdown from a Roon monolithic install on my i7/Radeon iMac. The cymbals seemed muted with that configuration. Countdown is a record noted for its clean, extended cymbal sound.
I’ll dig out the disk and play it on the Cambridge 851 using its internal DAC and unbalanced outputs. This is usually pretty pristine. I’ve not checked to see whose conversion bits are in the Cambridge. They do license 3rd party upsampling from a Swiss consultancy.
Just wanted to put my tuppenceworth in. Due to various work arrangements today I’ve been flitting between listening to Roon via. MacBook Pro USB output into a Chord Mojo and via. an Allo Digione (via. WiFi) into the Mojo. There is an audible difference in sound quality. Whether one is better than the other is sure to be a point of debate as we all have preferences however, for me, the Allo/Mojo has a warmer more natural and ‘listenable’ sound. Hence the recommendation to keep server/core/remote separate appears to hold in my case.
It takes experience or training to hear relatively minor differences, and it requires the system setup and room to be reasonably good enough, with proper music selection (e.g. a symphony for certain types of differences) at a high volume. The founder of a certain manufacturer said that it had taken him several years to be able to differentiate brass vs aluminum cones under a CD player in blind test. With a certain controversial format that I don’t want to name, many people with lesser systems said they could not hear any difference. In another forum where there are more people with expensive systems, they do hear more differences - good or bad.
It might take me more than a few tokes to hear such a difference. 
5 Likes
And then stating the difference is night and day
1 Like
So in your Lumin ethernet system, do you hear differences between Roon Server vs. Roon? Yes or no?
my recommendation get it off your iMac onto an external music server running roon core. Its due to a host of tech advantages that have single handedly been the biggest improvement in my chain resulting in better listening experiences.
Do you mean feature wise or SQ wise?
I’m still not getting how the type of computer makes a difference sonically if your computer is in another room and another circuit, while the audio system is on an isolation transformer that further separates the power. With this much separation electrically and physically, and your equipment being designed at least with some competency, I don’t see how the type of computer or the use of Roon vs. Roon Server would make any difference. Of course, unless one computer is very slow. That’s another story.
As far as ethernet goes, I’ve mentioned that in my post above, how you can work through options 1-4.
Clearly, I’m not getting something that the rest seem to be, or I’m interpreting things wrong?
Sq wise much better. I use antipodes audio ds core server using roon. Audioquest diamond USB into chord tt HD port. Xlr out to gsx mk2 4 pin xlr balanced to hd800S.
There is proprietary tech hardware that benefit from lower noise from an external system that’s power is odapi. SSD drives that can run Flac, dsd, whatever… go to the antipode site to read up…
Well, you’re using USB and not ethernet? Yes, we’ve already discussed that if using USB connection, then all sorts of things come into play and the quality of the computer running Roon Core makes a difference.
Everyone’s computer will play differently with their DAC via USB. I’m only interested in ethernet connections between separate devices running Roon Core and Output devices.