Reach maximum server sound quality / transparency

Hello! How do you like the sound after the new Roon 1.7 update (build 521)? Literally from the first notes I hear that the sound has become a bit more transparent and sharper in detail.

I would dare call it a little bit more identical to all the previous releases. Or, I daresay, a bit perfect, as it should be, is, and was.

8 Likes

“More identical”? Somewhat unique, that.

1 Like

Nicolas, you are all about right.
No one adresses your wish so far…so may I start.
First…at least for me it’s a big step forward in transparency what build 521 does. The subjectivity of recognizing this is due to how close your systems transparency does get to the treshold of being able to hear it (the treshold of anyones subjective ears/brain/intention/will are other factors ;-).
For Roon:

  1. Go through Settings and turn off everything monitoring or doing automated stuff in the background that you can afford not to have for comfort. Everthing counts, everything produces noise thus produces less transparency
    For example… Turn of all monitoring under Library. Only turn them two on if you add new albums to your library. Made a big deal for me.
  2. I don’t know your system but I use the following software: AO 3.0, Fidelizer Pro, Process Lasso on my computer. If you get PL you can direct all audio processes to one single core and give them highest priority etc…you’ve to fiddle a bit with PL on this.
    One important thing that did all the difference in my system was to direct ‘processreaper’ (process of Roon) to the core you directed all the audio processes to. That was a big one in my system. Try to avoid switching cores with all your audio software processes.

gui

2 Likes

I have my computer with Roon core 100% galvanic isolated from my HiFi, so I don’t even hear any decrease in sound quality when playing games on it, or running benchmarks unless its so heavy that Roon starts to skip.

Obviously Fidelizer etc makes no difference then.

2 Likes

I find it identical to every other release and identical to every other piece of music playing software connected to my DAC. Because surprise surprise, they’re all bit perfect.

I should keep links when I find them, but a few years back someone analysed the digital signal coming out of iTunes, amara, and Winamp etc.

The results were that all output waveforms compared were utterly identical in every single way.

I don’t know what you think roon does in the background that changes the way audio is perfectly relayed to the output with no changes, but I’m fairly sure roon have made no changes to the audio output part of their builds for some time, suggesting that any improvement you hear is purely imagined.

Worrying what your mind can trick you into hearing huh?

My advice, get over it. All music playback software sounds the same, all usb cables sound the same and all network cables, switches and routers pass the same bits of data whether they have an audiophile price tag or not.

You’ll save yourself a lot of grief when you realise that.

9 Likes

David, the people who worry about this seem to really enjoy worrying about it. So I don’t think you’re doing them a favor by trying to educate them.

4 Likes

it seems odd to enjoy worrying about something that doesn’t exist. I’d almost be inclined to suggest its a mild form of insanity. But if it makes them happy (which oddly it doesn’t seem to, as they never quite get to the level of “transparency” they seek)

life is very disappointing when you have unobtainable objectives

4 Likes

gui, thanks for the good idea of setting up - disabling background library update.

There are so many variables to tweak in file-based playback, it’s difficult to share the conclusion that RAAT sounds softer (or any other quality) than other architecture. It’s even more difficult to share that conclusion if it takes a $100k system to hear it. How many Roon devs have systems that resolving?

Some variables that I think are worth exploring include the hardware (Pink Faun, DIY server and/or endpoint with low latency, reduced noise hardware); operating system (Audiolinux, Euphony); power supply solutions (I’m sure you’ve experimented with this, but it’s worth mentioning); and network solutions (bridged connection between server and endpoint, re-clocked network switch). I just don’t think you can confidently attribute transparency problems to RAAT or Roon without eliminating those factors.

There are just too many variables, when we get to a $100k system, that it’s just not going to be compelling for the folks at Roon to start investigating those sound quality claims. Why tweak RAAT a little bit (or a lot), if those gains are going to be masked by going from a bridged network connection between server and endpoint to a switched connection? Doesn’t make sense, but you have my vote anyway. +1

1 Like

Perhaps I missed it, but please describe your Roon system in detail: Core, Outputs, Displays, and Controls. Also describe how each of these components is connected to your home network (copper/fiber Ethernet, Wi-Fi) and how each DAC (if separate) is connected to each Output.

There are a lot of smart people here who can help you spot something that you may have overlooked in your set up of Roon. No matter how experienced we are, it’s easy to miss something in the computer audio world since evey thing changes so fast and (almost) everything matters. Past knowledge is not always applicable and can lead to suboptimal configurations.

Leverage the knowledge of this group.

@Serhii_Kaidan You’re wellcome.
I’m in this business for over 30y. It’s kind of annoying to ever hear again the 'bit perfect" argument. For me it’s a closed minded and full of fear argument. Who will tell if ‘bit perfect’ is the one and only parameter for equality of
digital sound? The same goes with galvanic isolation. No question galvanic isolation in some places is a good thing but is it the one and only that gives for good sound on a music server? There might be other things to be taken into consideration.
Please don’t let us be so closed minded. For me it’s a faszinating journey to understand the origins of sound. Every day there’s something new…stay curious.

gui

2 Likes

And I realize that you wrote this post to spur Roon towards better sound quality rather than to request help debugging your system. However, I’ve talked to hundreds of Roon users and most of them have ignored Roon Labs’ recommendations for best sound quality to varying degrees.

Perhaps you have done everything perfectly; however, it’s worth confirming. So many who express dissatisfaction with Roon’s sound quality have not read and followed the manufacturer’s instructions.

1 Like

Fear? The opposite is true. It is a big relief for me that an ethernet connection transfers data bit-perfect.
So I don’t have to fear that audio quality is not as good as it could be and also don’t have to fear that my reports I write in word will change, our my online banking will do crazy things because the bits are not ‘transparent’ enough today.

7 Likes

There are at least 2 things that affect the sound besides the bits in digital audio:

  1. Electronic noise
  2. Jitter (mainly for SPDIF, but also to lesser extent clock-jitter)

According to John Swenson and some other guys, you also have something called phase noise, but I don’t know enough about it to agree or disagree.

Music-playing software (like Roon) don’t affect jitter, at least I don’t think so, but it will affect electronic noise unless you have totally isolated your computer, because more CPU, processes, GFX etc leads to more noise from computer. Why do you think Roon suggests separating core and endpoint? Are those guys also imagining sound improvements?

Its actually a little amazing that people still stick to the “bits-are-bits” argument, I though the human race left those beliefs 10 years ago.

2 Likes

You really want to have nightmares about jitter, think about streaming over Starlink. About 4 minutes into this video, Mark Handley starts talking about jitter of 1 or 2 milliseconds!

[Updated] Sorry, that’s Mark Handley, not David Handley.

Tell me one thing. Why did the sound quality didn’t change if I remove the connection cable from my computer?
My system plays along from buffer for maybe one minute after disconnecting. And there is not the smallest change in the sound.
Please explain how my cheap not isolated not audiophile enhanced computer, which runs a lot of parallel processes (it is also my work station) can do such a miracle. Shouldn’t it “affect electronic noise unless you have totally isolated your computer”, that I don’t have.

2 Likes

This.

We still don’t understand how to deal with all the electronic noise in the computer environment that is inescapable in file-based playback. Look at what Euphony is doing with CPU core regulation. Look at what Pink Faun is doing with their custom OCXO’s. I went from ROCK on a NUC to Audiolinux & Roon bridge on the NUC (with the server on a separate machine on the same network), and the quality improvement was significant. Both setups had the NUC connected to the DAC via USB. The NUC went from good, but with a harshness that caused listening fatigue, to clean and non-fatiguing. The things people are saying about Euphony are also really compelling. Those are just operating systems!

I can’t speak intelligently about what anyone hears other than me. For what it’s worth, these are the things I’ve observed, in my system. I didn’t want USB to sound better than the PS Audio Bridge II. I had no interest in dicking around with esoteric Linux installs. I heard what I heard.

Part of the problem of having discussions about things like this is that, even if the conclusions being reached are universal, that doesn’t take the specific system context out of the discussion. For example, maybe if I spend $4k on a custom PC running Euphony or Audiolinux and Roon, with the individual components powered by independent rails of a custom power supply, etc. etc., and that improves the sound, the improvement will probably not be as severe as if I spent the same money on an amp upgrade. Such is my system. Yeah, maybe your system is such that upgrades in the server end or the renderer end won’t be audible. Then the answer is, no, don’t make that upgrade. Prioritization is critical in system building.

3 Likes

Noise is real and depending on whats in the audio chain taking steps to eliminate same can/does make a difference. I have first hand experience of that - static electricity and changing power source on a SOC acting as a LMS endpoint. However, bits are bits - the audio stream either arrives at the DAC intact or it doesn’t, and when it doesn’t the effects are immediately audible. Noise, however, doesn’t need to corrupt the bits to become audible. Perhaps a way to think of it is a multi-lane highway, bits traveling in one lane, noise in another. Both make it to the DAC. It’s how the DAC handles that, that matters most. For older DACs it certainly matters. For newer DACs one would hope it matters less - but I would think that very much depends on the design.

So, in my setup I know that powering an Odroid C2 (that served as a USB based LMS and later Roon endpoint) with a regulated power supply that does a great job dealing with incoming noise and generating very little of its own makes for a lower noise floor when paired with a particular DAC which I still have.

Whether it now makes a difference powering a Pi4 doing the same job but feeding a Yggy I have no idea. Why - 1) because it now feeds a Yellowtec PUC2 USB > AES/EBU converter which feeds the Yggy, 2) I couldn’t personally be bothered with further experimenting, the regulated power supply is still in the mix.

I guess I could add the Odroid C2 to the mix again, feed the Yggy’s USB directly off it and try compare perception of noise floor with and without the regulated power supply, but truth is I’d rather just enjoy the music.

In any event, beyond that …

sucker…every…minute…

My solution to that problem was to isolate my HiFi

  • My poweramps uses a dedicated power cable from a dedicated fuse
  • I use fiber between ethernet switch and streamer (2 * fiber media converters)
  • Media converter, streamer and DAC runs on battery
  • I cut pin1 on XLR on DAC side to severe the ground connection

500W power amps are a little difficulty to run on battery, but the rest is very isolated both from mains grid and from Roon core.

1 Like