Is there any sound quality improvement with the new Roon 1.8? What are the improvements, if any?

System jitter is in fact a real thing, and it could conceivably manifest in a purely software system with nondeterministic inputs. But I can’t see how it could possibly affect SQ.

Perhaps the problem is that, while Paul MacGowan knows what he is talking about, his listeners sometimes don’t. What he’s talking about is selling more equipment to gullible consumers. If his listeners mistakenly think he’s talking about what he says he’s talking about, which isn’t really what he’s talking about, misunderstandings can develop.

11 Likes

I rather suspect that’s a made up term. But I wholly grant that due to the way cpus “multitask” that if your endpoint is doing a lot of extra work then it could lead to playback issues if it is resource constrained.

Luckily Roons server/client architecture makes this a non issue, as all the work is done on the core and the client does next to nothing.

1 Like

Cool… I should have made clear that I tried Roon and Squeeze Server on the same Antipodes box – it runs several different players. I made the comparison several times over a number of weeks, and was, in fact, biased toward wanting Roon to sound better. I had been using it for a couple of years, and obviously nothing compares to its UI.

1 Like

I would like to see an option to reverse absolute phase. Since we already have DSP, maybe this would not be difficult to include. On many recordings this makes a significant difference. I am able to do this with my PS Audio DAC (built in) but not on my PC. I am sure there are many with DACs which do not have phase inversion as an option and might appreciate this being implemented in Roon. Since my first CD player many years ago that had it (Audio Alchemy) , I learned to appreciate it and would not buy a digital source, CD player or DAC that did not provide it.

Do a search on here. There’s a way to set DSP to invert phase. Once it’s set up you can save the settings and use it anytime you want.

I didn’t make up the term. I can’t take credit for it.

It has nothing to do with bit perfect playback. Ones are 1s and Zeros are 0s for sure.

Many get hung up on that for good reason.

I few years ago it was relatively to discern that one was playing a Flac file vs AIF vs WAV etc. The reason I’ve been told, is because the Flac file had to be opened or unwrapped. Those are not my terms either. One could hear this on good systems. Again not a bit perfect problem. Both played 1s and Os perfectly. And if you never compared the formats you’d never know the difference.

It’s not like you could walk into a room and declare, “Now that’s a Flac file playing. “

But if you took the time to compare the difference was there.

Now with today’s much faster processors that little difference is much less if detectable at all.

The Taiko Supreme Server uses two Xeon processors which is one reason it is believed to sound so good. Some of the guys at Audiophile Style have built PCs with two processors and have experienced the same uptick in sound quality.

I would hypothesis that as processor speeds increase the amount of “software jitter” will decrease similarly to the FLAC phenomena.

Who knows we may see servers built with three CPUs?

I love using ROON. I’m a lifetime guy. I look forward to experiencing the new look and feel.
But I hope they did a bit of coding to tweak the sound quality as well as the Valence.

The computer with the Core impacts the sound quality of the Server.

Just try adding Fidelizer or similar program to your Core pc. You can try it for free. Basically Fidelizer turns off parts of Windows do the processor can have less to do.

Fidelizer is free to try. I’d guess one could hear the difference with just about any level of system using a pc for playback.

Again this has nothing to do with bit perfect. It’s the extra processes on your PC that cause the perceived difference.

Ho Hum.
Believe what you wish.

I’m not hear to lie or sell any snake oil.

Technically all modern computers, even phones already have multiple cpus, between 2-4 actual cores is common and some have 8 or more, let’s not get into hyper threading.

And sure, because flac is compressed it needs to be uncompressed (unwrapped if you like) before it can be played and if your playback device had very limited resources and had to do this while playing it could cause noise. But roon does this on the core and sends uncompressed audio to the endpoint. In fact it would send the exact same bits for all three file sources (assuming they were the same recording/sample rate).

2 Likes

I see you removed my perfectly innocent post yesterday on the Roon 1.8 thread about SQ. Why? Are you so frightened when somebody asks a perfectly benign question about anticipated?

Hello @Martin_Wax, is this the post?

It’s there. Was there another one? I can’t see one.

Moved non-Support question to Roon Software category.

How can sound quality change when it is already bit perfect delivery ?

11 Likes

there is nothing to add…

1 Like

Yes. They have multi cores. Of course they do. And from that standpoint you’d think that would end the story.

I’m not sure how many core a Xeon chip has that they use in the Taiko Extreme but the designer uses two Xeons. Several of the folks at Audiophile Style have built PCs with two Xeons and they confirm the uptick in sound quality. I’ve not heard one in my system I still use a lowly i7 but those with i9s claim they are better two. I’m not going to say they are because I’ve not heard this demonstrated my self.

But when I decide to build a new server I’m building it with two Xeons.

From what I understand manufactures of today’s best streamer/server products design their next generation equipment around, faster processors.

Why would do that? Why of course so they won’t be outdone by their competitors.

You’ve hit the nail on the head there. They do it to be competitive. Whilst in reality an average computer from 2012 has more than enough power, consumers see product A has 20 thousand giga flops and product B has 30 thousand giga flops, B is clearly better. What’s a giga flop, who cares (it is actually a thing), more is always better.

I’m not trying to be difficult, but you could save yourself a lot of anxiety and money by not being taken in by such nonsense. If you have money to burn, then there are a lot of good charities that could benefit.

Check out https://www.givewell.org/charities/top-charities

3 Likes

Hi hum. I knew I shouldn’t even try.

I’m a teacher. Once in a while I make the mistake and try to teach.

The one thing I don’t understand is why some have to find the need to write about something that the already know the answers too.

That’s why I post less and less on these forums.

ROON users have a substantial interest in the software evolving both in perceived usability and yes in sound quality as well.

Personally I liked the old format just fine. It did everything I want it too and more probably than I’ll ever discover. It looks interesting for sure and I like Valence.

I just hope that the Wizards of ROON improve the perceived sound quality as well. If ever there was a programming team that can do both it’s the Wizards of ROON. (My term not theirs)

Merry listening to all and to all a good night.

3 Likes

For best SQ results Roon recommends separating Core and Output and using a network Ethernet connection between them. This enables a fair degree of isolation between the “noisy” Core and a low footprint Output which feeds the rest of your system. If more isolation is sought then optical Ethernet can be used.

Changing to this topology should make your system less prone to SQ variation between releases because, in the absence of any explicit SQ updates, such changes are likely to arise from more or less intensive computing in the Core.

In short, isolating your Core will do more for SQ than which version of Roon you run.

7 Likes

If you’re using Roon, and only Roon, on a dedicated Core then don’t waste your money. Roon cannot take advantage of such architecture. You’ll just be increasing your power bill for no reason.

The only advantage Xeon gives you, over the consumer processors, is ECC memory but I’ve seen no one claim this gives a sound quality improvement. The Taiko audio box is, I’m sorry, a foolish design to use for Roon. The Roon software is not optimized to do anything special with the hardware. Additionally, the cost of parts is significantly overinflated to what they offer. It’s your money… if it sings to your ears then go forth and bask in the ownership. Those of us who have a basic understanding of software, CPU, memory, and bus architecture will happily spend our money where it matters; speakers, analog electronics, and room acoustics.

Disclaimer: I run Roon as a VM on a Xeon server with ECC memory. It sits in a closet on the other side of the house from my main system. I’ve never had performance problems. Sound quality has been improved with better electronics in the main listening room and optimizing speaker placement and room treatment with every change. When my server dies I’ll replace it. Until then, I’ll keep improving the stuff that sits on the analog side of the system as I believe that is the better use of my cash and provides significant, and measurable (I use REW), performance increase.

9 Likes

Comment on measurements and filters / upsampling…

One of the main reasons you won’t see measurements for some of this stuff is because the measurements only show you what changed. What changed doesn’t equate to better “sound quality”. In digital to analog conversion there are artifacts and errors. There is never a perfect capture of the analog and there is never a perfect playback from D to A. Those errors and artifacts affect our “feeling” on how something sounds. And, we’re all affected differently. Not everyone agrees vanilla is the best ice creme (edit: or how to spell ice cream apparently. I owning it though. Thanks @Bill_Janssen).

Go look at these examples of what happens when you apply the different filters to Roon’s upsampling DSP https://help.roonlabs.com/portal/en/kb/articles/dsp-engine-sample-rate-conversion

What ones right? What measurement is better?
That’s why you don’t see measurements but someone can still say something sounds better.

I recently did a comparison of the filters in my DAC vs. Roon… I liked my DAC. But in another room, different DAC, I prefer Roon DSP. I used to dial-in my own isotope SRC settings in Audirvana. Glorious sound! Maybe even better than Roon. I know I’d probably see improvement with HQPlayer. But… Roon is, essentially, the truth and all those other things are just different ways of letting me swap-in / out different errors and artifacts until I find the most pleasing sound for me. If I measured that it’d only be a perfect measurement for me.

Anywhooo… so… yes, people will hear improvements with different software, filters, etc. Yes, you could measure this. But, one persons “better sound quality” is another person looking to get a refund. I like the fact that Roon presents, to my ears at least, the truth. This allows for everything around Roon to be exceedingly transparent. Exceedingly transparent can sound bad unfortunately.

2 Likes