A very quick summary of my findings on my main system.
These days I simply use Roon for all of my music listening without even considering the possibility that I might be losing out on the last word in sound quality by doing so. It’s been quite a few years since I last compared sound quality from Roon versus other streaming platforms, but this thread has prompted me to give it another go.
I think I have a reasonably good main system comprising Linn Klimax (DS/1) streaming DAC, Lindemann 830S pre-amp, Lindemann dual mono power amp and Magnepan 3.6R speakers. I subscribe to both Tidal and Qobuz, but for the purpose of todays tests I used tracks from Qobuz via Linn Kazoo (Linn’s own software) vs the same tracks from Qobuz via Roon - all tracks were standard ‘CD’ quality (44.1 kHz, 16 bit). I also tested CD quality tracks from my local NAS using Roon vs the same tracks streamed from Qobuz using Roon.
My focus when testing was based on determining whether or not I could hear a difference of any sort between different delivery paths. My subjective preference of one delivery method versus another could be introduced later on if necessary.
In summary, I could not detect any discernible difference between tracks delivered from Qobuz by Linn Kazoo and the same tracks delivered from Qobuz by Roon, nor from tracks delivered by Roon from Qobuz and tracks delivered by Roon from my local NAS.
To be completely accurate, very occasionally there were times when I did wonder if I could actually hear very subtle differences. However, the differences were so hugely subtle (and not at all consistent) that I would attribute them to deficiencies in my short term audio memory and my testing process rather than differences in actual sound quality.
Of course, these are just my personal findings with my audio system setup and I don’t dismiss the possibility that others with different audio equipment setups and possibly younger ears than mine may genuinely hear differences.
Now, I haven’t actually tested Qobuz direct vs Qobuz via Roon. However, the tests I have run to compare Kazoo vs Roon have reassured me that I do not need to worry about any loss in sound quality when I use Roon on my systems.
My listening setup in the living room is as follows.
My Cambridge Audio CXNv2 streamer is hooked up digitally (optical) to my Marantz M-CR612. The Marantz passes through a maximum of 24 bits / 192 kHz this way. The Marantz drives a pair of Dali Spektor 2 speakers and a JVC subwoofer.
I couldn’t test Qobuz directly on the Marantz as the HEOS app doesn’t support Qobuz.
So what I’ve done is I’ve tested Qobuz on the Cambridge through Roon and StreamMagic. StreamMagic is Cambridge’s own app. Next to that I’ve been using the Qobuz app on my old Nokia phone, which I’m using as remote in the living room. The Nokia runs on Android and the Qobuz app connected to my Cambridge through Chromecast, maxing out at 24 bits / 96 kHz.
I’ve been testing one of my favorite tracks. Donna Summer’s “I Feel Love” as the 12" version, which Qobuz hosts as 24 bits / 192 kHz.
I couldn’t make out any differences in sound quality whatsoever.
Interesting, thank you. I’m not at all familiar with Kazoo but googling it implies that it isn’t a step in the signal chain but rather it’s Linn’s controller, rather like StreamMagic is Cambridge Audio’s - they are richly featured phone or tablet or desktop based remote controls that tell your streamer/DAC what to do and which signal to use, but they don’t form a part of the chain through which that signal goes.
If I am right, and assuming that your observations of differential sound quality are as valid as those of us who are seeing a difference between “Qobuz direct to streamer” and “Qobuz via ROON” it would seem to give credence to what I have been suspecting for a while now, which is that ROON doesn’t degrade the signal but that the signal that arrives at a DAC (whether its built into a combined streamer/DAC or not) is pretty much the same whether it comes direct from Qobuz or whether it comes via ROON.
Which would imply that some systems, on receiving these signals, put them through a different processing path.
I’m no sound engineer but if on some really pretty good systems Qobuz direct sounds the same as Qobuz via ROON then in other systems where the two sound different, the only distinguishing feature is the signal path in that system. No?
This has also been my experience (two Linn Klimax systems), but note that the Roon core>Linn DSM connection uses Linn’s own multi-endpoint protocol stack (Linn Songcast, called “Linn streaming” by Roon), not Roon’s RAAT. That means that there will be fewer code and data path differences between sources in Linn gear than with other brands that implemented Roon RAAT starting rom Roon’s reference implementation. Having worked in the software business in various technical and managerial capacities for decades, I can tell you that “it should behave exactly the same, but it does not” is one of the most frustratingly hard situations to debug, especially when the evidence of difference is subjective or intermittent.
A couple of comments: firstly, Stream Magic isn’t a ‘through’ scenario, it’s just a remote control that doesn’t process the signal at all. ROON actually is a ‘through’ scenario in that it gets the stream from Qobuz and does stuff to it before streaming it to your system. Secondly, though I have no experience of the JVC subwoofer I do own the Dali 2 in a second home and great though they are for the size and price I wouldn’t think of them as very revealing of differences at this level. I could bring them to home #1 at some point and test that hypothesis by wiring my EVO to speakers channel 2 and switching back and forwards but it will take some time before I can do that.
As most here probably know, the human auditory system is hugely fallible (we’re primarily visual creatures). Given that, instead of uncontrolled listening tests that are subject to varying amounts of bias (my own included), I find it more telling to conduct null tests. If you’re unfamiliar with audio null testing, there are lots of references on the web.
I was curious about the OP’s question so I performed a null test comparing a Qobuz stream from the Qobuz app, and a Qobuz stream through Roon. Both snippets were ripped using AudioHijack and compared in Ableton. I lined up the samples at the sample level and reversed the phase on one of the snippets. The resulting third track in the image is the difference between the two source samples. What is there is low level and consists of mostly unintelligible high frequency clicks and noise. This could be a very small legitimate difference, or more likely it’s due to slight clock drift. In either case, this level of difference is unlikely to be audible.
I hope this is helpful. I’m anticipating that some people will instruct me to trust my ears, but when it comes to differences as insignificant as this, I like to back up what I believe I’m hearing with something more tangible.
I’m sure what you say is true but it doesn’t allow for what our various components do once the stream arrives with them from either Qobuz direct or from ROON via RAAT. The difference I hear is not particularly subtle. It’s not egregious to the point that one would say ‘the ROON stream sounds bad’ but it’s pretty obvious to me that on my system it does not sound as good. It’s as obvious as when you go to the HiFi dealer for a demo and select one component over another because the sound is obviously better in an A/B test.
In my opinion, the most important thing is to discern whether the raw streams are different. Once that piece is understood, the downstream effects should be relatively easy to figure out. In my testing, the differences are at such a low level as to be inaudible, so I’d be looking elsewhere for the cause.
I can hear the same differences and with the same ease on both my $10k system as well as the $2kish system (Emotiva DC-1 and HR 824 mk1…I do have a good room though). At this point, I don’t know if the difference is actually that big or I’ve just trained my ears/brain with an obscene and unhealthy amount of time doing testing for myself and others in beta testing. But yes, the difference can seem like a change in components type of level. I really wished they all sounded the same. The dryness and lack of ‘sweetness/musicality’ is hard to ignore when switching over.
Respectfully I am using a 3k+ pair of speakers and the Dalis are about 0.4k. If your point is that any difference between speakers at those points of price and ability isn’t significant to your ears then I understand that - it’s your choice - but if you mean to imply that the differences are so recherché that they practically amount to nothing in a more widespread sense, you’re making the case that no one should bother spending more than about 400 on a pair of speakers - and I don’t think many people would agree with that.
@SandsOfArrakis I must say that Alan’s experiment does not address the issue fully. It doesn’t look at the possibility that some streamer/DAC/Preamp units might be treating the stream they receive from ROON differently from the stream they get direct from Qobuz. I don’t know. But I do know that the more you spend on HiFi, ceteris paribus, the better the sound. Probably with diminishing returns at some point. But the idea that a $400 pair of bookshelf speakers will be as revealing as a $3K+ pair of stand mounts would really imply that everyone with gear above low mid-range has been conned…. It’s not a matter for democratic outrage that more expensive things are better, they often just are. No one is saying that your opinion is invalid, just that it cannot be fully informed until you have tried the alternatives - and as you will note from the above, I also own the Dali Spektor 2 and they are very good for the money but they are not capable of what the ATC SMC19s are, by a very long shot. You wouldn’t expect a McDonalds to taste as refined as a wagyu steak, however much you might enjoy it personally and find the accessibility of its price point more acceptable.
Unfortunately, in the world of HiFi there is clear correlation between price and sound quality. For some components (cough, cables, cough) whether you pay X or 100X makes no difference whatsoever. For speakers… yes, there’s a decent chance that a $10K speaker will sound better than a $400 one. There’s a pretty good chance that it will sound worse, too.
