No, that’s of course not true - even COMPLETELY MYTHOLOGICAL!
It’s a strictly “analog” thinking.
Unless it’s a provocation?
I have very old “self builded” PCs but, properly configured, and roon sounds absolutely fantastic, easily reference standard - for me roon (with Roon Bridge, using Roon DSP) sounds better then others I use from time to time.
My ears are of the same opinion. 
Theoretically speaking, I can imagine that there is something to this. Practically speaking, however, I’m convinced it would be extremely difficult (if not impossible) to design and conduct a listening test that reliably verifies (or falsifies) such a hypothesis.
Although I have no idea what rubbish sounds like, it appears to me that it would be quite an understatement to call this statement a bit of an overstatement. 
The audiophile world is mythological by it’s nature 
My lack of knowledge of mythology has always been my achilles elbow…
This is wholly inaccurate. Roon doesn’t have a “sound”: It reads data from a hard disk or network streaming source and sends it to a destination end-point over a network as TCP packets. And using Roon, versus some other “audiophile” player is NOT analogous to swapping $5000 tonearms on an overpriced turntable.
Therefore, the more money spent on the computer, the better it will sound? How, and why? And why are evaluations being made against “high end sources”, all of which appear to me to be insanely overpriced computers using off-the-shelf components, enclosed in six inches of machined aluminium to satisfy the vanity of analogue-era audiophiles that cannot help but equate price with quality?
You “must” do no such thing; that’s why they’re called “recommendations”: They exist to help Roon with customer support issues more than anything else. You very much can “cobble together” old hardware and have it work equally well as a $20,000 “high end source”. That 10 year-old PC in your basement with a first-generation Core i7? Load it with Ubuntu Server and you have a perfect Roon Core. I’ve done it and there’s absolutely nothing I can do in Roon that will have it using more than 50% of its CPU resources. $60 Raspberry Pi’s make perfect end-points that never need to be touched: Your microRendu is the overpriced end-point equivalent to the $20,000 “appliance” and makes all the same ludicrous claims to sonic perfection.
This must come as quite a surprise to those, like myself, that use them as end-points every day and can discern no degradation in sound quality. And what a shame Roon wasted so much time porting their application to run as a Core on those platforms!
I use drivers from roon specific to by d/a converter and the sounds improves substantially.
What is your complete chain, from Roon Core to speakers/headphones, if you don’t mind me asking?
In this thread, in which end users are discussing how they perceive the sound quality of Roon in comparison with other audio players, the (correct[!]) statement that “Roon doesn’t have a ‘sound’“ could easily be misinterpreted. For many end users the decisive question is how the way Roon “reads data” and “sends it to a destination end-point” influences the sound quality they achieve with their system. Against this background, I think such linguistic inaccuracies should be regarded as ‘tolerable‘.
are you running the right bit-rate - 32 or 64? Are you running the ios download on a windows platform or vice versa? Easy to click on the wrong one.
My Roon Core system is an Intel Core i5 3570K (Ivy Bridge) system (with an ASUS motherboard whose model I can’t recall) running Ubuntu Server. I originally put it together as a 16TB NAS and Plex server, and it still serves those functions alongside Roon. It has 16GB of memory and an SSD boot disk, but all data, including Roon’s library, is on traditional 3.5" Western Digital Caviar Green disks configured as a striped array. It’s in a Fractal Design Define R5 case which has some acoustic dampening, but it lives in the furnace room in the basement where I can’t hear it.
Note that this computer is approaching its 6-year anniversary and I don’t anticipate needing to replace or upgrade it until something critical like the motherboard dies and I’ll be forced to seek a new processor because of constantly-changing socket architectures. Or my brother upgrades his computer and gives me his old motherboard & CPU, which is where the ones I use now came from.
I have Raspberry Pi 3 end-points in my office, living room and bedroom that connect back to the Core with gigabit Ethernet (though the RPi’s are limited to 100Mbps). I have five or six Ethernet switches throughout the house, but they all aggregate to a central switch in the office where the fibre Internet connection enters the house.
Office RPi:
- connected via USB to a Schiit Modi 2 Uber, which in turn connects to a NAD C326BEE integrated amplifier and a pair of PSB Image 1B speakers
- connected via USB to a Schiit Bifrost Multibit, which connects to a Schiit Jotunheim headphone amplifier
Living Room RPi:
- connected via USB to a Schiit Gungnir Multibit, which connects to a Schiit Freya pre-amp along with a turntable. This setup drives a pair of PSB Image 6T speakers through the amplifier section of a Bryston B60 integrated amplifier
Bedroom RPi:
- connected via USB to a Schiit Modi Multibit, which connects to a Schiit Asgard 2 headphone amp.
Headphones are Sennheiser HD 600 & HD 650, NAD VISO HP50, PSB M4U-1 and Bowers & Wilkins P5 & P7’s.
I use Straight Wire ‘Symphony II’ analog interconnects because they’re inexpensive, well-built and not ugly. USB cables are all AmazonBasics: Cheap, well-built, flexible and most importantly, black. Ethernet cables are either whatever I had on-hand to suit the length, or cut and terminated by myself from a spool of StarTech’s generic Cat-5e. Speakers cables were all made by myself with Monoprice 102747 12AWG cable and Monoprice banana plugs.
Not terribly difficult to test.
However, to save anyone that trouble, here would be the results of those tests. Both digital bit streams would be bit perfect identical. And both sets of analog output measurements would be identical – within experimental margin of error that comes from measuring analog signals separated in time.
Any listening tests would be superfluous. And any perceived differences would be imagined.
Brian already posted this link, but to see a related demonstration of what the above tests and their results would look like, here is the link again:
AJ
In the post you replied to I explicitly wrote:
Much to my surprise, your (provocatively brief?) response to this was:
Against this background, I must confess I was “terribly” disappointed (only kidding!) when you suggested the following test method:
Quite obviously, this is NOT a listening test. Of course, I quite agree with you that your informal experiment would be fairly easy to carry out, and I also would expect the same results as you (i.e. nearly identical analog output measurements). However, I couldn’t disagree more with your statement that
As I pointed out in two earlier posts, I’m fully aware that (informal/private) listening tests “always have to be regarded with a healthy dose of scepticism” and that “anyone can fall prey to” what I called the “powerful effect of preconceived notions” (see above). Yet at the same time there can be no doubt that – together with the most accurate/precise measurements possible — carefully/scientifically conducted listening tests are an indispensable (additional) tool in this particular field of research (even though they can be extremely difficult to construct and perform). Any kind of ‘either/or’ thinking is clearly counterproductive in this context. In fact, it appears to me that the often-heard assumption that “if you can’t measure it, then it doesn’t exist” is (almost, but not quite ) as naive and unscientific as the quasi-religious belief of a small minority of self-proclaimed audiophile “experts” who seem to think that the world of HiFi does not follow the laws of physics and that their “golden ears” render any measurements superfluous.
Some folks rightly figured out my post was meant to liven up the discussion a little bit and thanks to everyone who read/replied. But seriously, to really enjoy hi-end, it helps to avoid preconceived notions based on “science”. The history of hifi is full of improvements that defied the science of the day. For example, in the 70’s people said speaker cable was a rip-off because it was double the price of lamp flex and measured the same. Today do you use lamp flex to wire your speakers? And CD’s gave you “perfect sound for ever” because bit are bits, right? A Roon digital audio reproduction chain is more complex than a length of wire or a CD player so why is it a surprise that the hardware can make an audible difference? I’ve run Roon core on Synology NAS, i5 and i7 and each step up there is an audible improvement. This was not a good finding as I wanted to run core on the DS918+ but now I have the hassle of an always-on quad core i7. Get some friends together and try it. If it makes no difference in your system you still had a fun afternoon. Regarding endpoint, if you have the technical knowledge to build a microrendu equivalent by replacing all the switching power supplies in your old ATX board with high performance linear and install a minimal linux that does nothing except run Roon enpoint then go ahead. Most folks should just consider a purpose designed end point - you are paying for a professional engineer’s time, not just a pile of metal and plastic. BTW I do agree Pi as endpoint is quite good and I have one but it’s not quite as good as microrendu.
These are honest questions: How did you arrive at this conclusion? How did you compare the three devices? Do you have as many as three Roon subscriptions or did you move the core from one device to the next and then try to “remember” what it sounded like “last time”?
I honestly don’t know how my friends and I could try this. If you do, please tell me (I don’t mean this ironically!).
Well said, really well…
Is this called sophistry?
Comparison between i5 and i7 by CPU swap, 30 minutes work, no other changes. I’m the kind of person who has a bunch of CPUs of the same series lying around. Previously with J River the i5 gave a better sound than the i7 presumably due to its lower power draw. So when I reformatted the PC for Roon, I kept the i5 and gave it no further thought. The only reason I put in the i7 was because I needed to run handbrake on that machine. I was certainly not looking for any change in the sound but I found myself enjoying the music more and eventually realised why. If someone is up for a scientific A/B please let us know your results. I think If folks are going to get a new computer to run core then going straight for an i7 makes sense. Anyone already running core on i5 should not worry about it too much until satisfied with end point / DAC / amps and looking for the next improvement. Or if you have a spare i7 lying around.
I sincerely fail to understand how processor speed would affect SQ. I just can’t think of any argument to support that. A processor is either doing it’s job or not. “Not” being, 100% utilized or not working at all. The data is either getting from the hard drive and memory to and from the software or it’s not. Again, “not” being characterized by delays due to over-utilization or “not at all” for some other reason, which I could of course see affecting SQ.
If the CPU is under 100%, it’s working. The bits are being processed and sent along to where they’re needed.
We know that the recommended processor depends on things like library size, use of DSP etc. But this isn’t a SQ question, it’s a question of whether the CPU can handle the greater processing demand, resulting in one of the two scenarios I outlined above. Not some mysterious degrading of SQ.
Just my $.02. Entirely possible there’s something I’m overlooking.
Of course would not - it’s obvious, that’s physics.
He’s just kidding around
oh. lol.
carry on then…