Sound quality - not as good as other programs?

Roon Software Discussion
21

Hi Everyone,

Thanks for the great advice so far.

Having been playing with Roon for the last 2 days, I think I have decided that the amazing UI and the fact that I keep discovering new music has convinced me to keep Roon over Lightning DS, even though I do believe DS has better sound quality. Perhaps its better by 10-20%, although difficult to quantify.

For clarity, I am using my MacBook Pro as the Roon Core, as I found the NAS really struggled (low CPU power). I am running the latest version of OSX.

So if anyone has further tips or tricks to improve SQ I would love to hear it. Hereā€™s my system:

Synology 916+ NAS
Asus RT-AC88u Router
Macbook Pro
Auralic Aries LE
Chord 2qute
Gieseler Linear PSU for Auralic and Chord
Marantz SR7008 AVR / Pre
Anthem MCA20
Focal Scala Utopia
Analysis Plus Solo Crystal I/C and spkr cable
Curiouscables USB (Aries to DAC)

Here is a summary of the differences I have observed:

I find the sound with Roon to be ā€œwarmā€ but not in a particularly pleasant way - seems the warmth is robbing the high end detail.

Dynamics - good with both, probably equal

Bass - Slight edge probably goes to roon here acutally. Seems to dig a little deeper

Transparency - definately better with DS - just seems more ā€œairyā€

Stage - Wider with DS, likely a result of the high end detail and subtle aural queues

Image - good with both, probably equal. DS is a little more 3 dimensional.

Sibilance - interestingly Roon seems to have slightly more sibilance despite having less overall detail. Cant explain this one

22

Might sound odd, might not; Iā€™ve found, using just one DAC, that playing with the Clock Master Priority setting, found under Audio setup -> Zone Grouping, had a subtle but noticeable influence on the sound. At the highest setting of 1 gave a little tighter sound, more focused, but became boring. Reducing to the lowest setting, 10, the sound became a little more relaxed, airy, less strained, but lost a little focus and became grainy. Just my observations using a Windows 10 PC.

23

So Iā€™ve had an interesting revolution.

Firstly, I realise that using an AVR isnt ideal for High end HiFi, but I plan to get a nice integrated soon, so the AVR is removed from the signal chain. In the interim, I use the 7.1ch analog inputs, so this ensures that the analog signal from my DAC doesnt get converted again. This is the the purest way that I can use the AVR as just a preamp. So no audyssey or anything, just straight analog in, analog out (via preout)

Anyway, I started by running Audyssey MultiEQ XT32 and then copied the settings (ie the peaks and dips by freq)

I then enabled DSP and Parametric EQ in roon and entered the exact same settings that Audyssey called for. I added a boost at 63hz as Audyssey cuts this by -6db - so I went with -3.

Well, I never thought I would say this, but wow, this EQ is fantastic!

I would now say, that with my EQ tinkering, the SQ is equal of that of Lightning. Yes I realise that the Purists will claim EQ is evil, however, for me it has helped get closer to the SQ I desire.

1 Like
24

Hi - What does the Clock Master Priority setting actually do? I agree, subtle differences between the different settings, but not sure exactly how/why?

25

I believe itā€™s used for when you have more than one zone/end point to control clock drift/sync issues with the other DAC devices. Donā€™t know why it affects the sound, no matter how subtle, but after more than a decade of building computer-based transports for DACs, small changes in hardware and software does affect jitter/clock timing and resultant sound quality.

26

Now, you can use a Linn Klimax Streamer as a Zoneplayer using Products from Sonoreā€¦
People say there is no degradation to their usual ā€œKlimaxsoundā€.
Or you can have a ā€œNaimsoundā€, using a Naimstreamer.
Or any DAC from ā‚¬100 to ā‚¬50000.
All over your network setup. No noise from any PC running software.
You can have the soundquality you like or you can afford.

27

this commercially or knocking together transports for your own use?

28

that be for my own use

29

So now Iā€™ve stumbled across the ā€œsample rate conversionā€ option in the DSP

Any thoughts on this? Iā€™m guessing it takes the original signal and upsamples it to whatever setting I use (Iā€™d use DSD128)

Had a brief play around and setting it to 7th order DSD128 made the sound a little smoother and cleaner (maybe a little more laid back) but seemed to have robbed a little detail/dynamics.

My Chord DAC upsamples anyway, so wondering if this is a good/bad idea to use?

30

you answered that rhetorical several times ā€œjust listenā€ trust your own ears.
I do use up-sample and i like it but we al have different systems so results will be different. So that renders the advice of many to nothing :slight_smile:
Goodluck and enjoy Roon

31

Roonā€™s high-quality upsampling works a treat with my non-oversampling DAC using the minimum phase filter. But maybe different for other DACs like the Chord because theyā€™ve put a lot of technology and money into itā€™s FPGA oversampling filter algorithms.

32

This is a real can of worms and there is information scattered across these and other forums like Computer Audiophile about Chord DACs internal upsampling, handling of DSD, (and Rob Watts thoughts on DSD in general) as well as the ideal upsampling (if any) settings for Chord DACs. Long story short the general recommendation is that its best to upsample to max rate PCM, if at all. No disrespect intended but upsampling in Roon with a Chord DAC strikes me as a bit like putting a saddle on a fine British motorcycle :slight_smile:

33

Assuming the perfect DAC (quite happy with my Chord 2Qute), if convolution is applied for room correction, would it matter if that is done at 44.1 kHz, or after (Roon) upsampling? Roon DSP is performed after upsampling, and the sample rate affects the number of taps.

34

There is no jitter with TCP/IP transport. It contains no timing information.

If you mean interface borne jitter between the transport and DAC, well I suppose so, but any competenly designed DAC should effectively reclock. The numbers are vanishingly small these days. Even under conditions where jitter figures have been artificially pumped to give people the best chance to hear a difference itā€™s been shown that they canā€™t do it under blind conditions.

1 Like
35

I have a 2qute too and upsample as a default to 384khz

To my ears definitely an improvement through my rig though not the DSD which loses too much top end sparkle no doubt about it.

PCM upsample into chord thumbs up from me and itā€™s fairly significant as an sq improvementā€¦ imho

2 Likes
36

Iā€™ve decided that with my Oppo HA-2SE the Roon upsampling and DSD conversion isnā€™t as good as just simply getting a bitperfect stream out so I have DSP turned off. I think HQPlayer does a better job with upsampling but I found it too buggy so Iā€™m not using it anymore.

37

I hesitate to jump into this thread, but I think that a lot of people involved in these A+ vs everything else threads are missing a critical point.

When you install A+ you also install his system optimizer which (among other things) makes a bunch of configuration changes to the computer every time you press play and then undoes them when you hit stop. This turns off background processes like Time Machine and Spotlight indexing and also does some re-jiggering of process and I/O priority. It essentially turns off as much of the ā€œnoiseā€ as it possibly can and ensures that A+ has higher priority than most other processes. This is great and itā€™s one of the reasons why A+ is a good playback solution for a computer directly connected to a DAC.

Jitter has become the default boogeyman in high-end audio because itā€™s just counter-intuitive enough to a be a mystery to a number of people. It sure sounds like a nasty thing, thoughā€¦

Jitter in and of itself is an issue, but in a properly-implemented system itā€™s almost irrelevant on its own. Noise is the killer and this can be as simple as a ground loop to internal reflections within a cable due to impedance mismatches. A+ does a better than average job with computer to DAC playback systems because it tries to extinguish as many sources of noise within the computer as possible.

Roon has consistently recommended that in order to optimize sound quality you need to isolate the core function from the playback function. This is either through the use of a Roon Ready device or via Roonā€™s own RoonBridge program. Iā€™ve gotten outstanding results using Roon into a Roon Ready DAC as well as with RoonBridge installed on a dedicated low-power linux box (Intel Atom).

Is this ā€œbetterā€ than A+? I donā€™t know, and really donā€™t care. Iā€™ve used this approach on very modestly-priced systems ($5K) up to those which could be considered all out assaults on the state-of-the-art ($500K). The sound quality is outstanding on every system on which Iā€™ve used this solution and when combined with Roonā€™s interface thereā€™s simply no comparison.

8 Likes
38

on the premise that Roonā€™s UI is so good I canā€™t use anything else anymoreā€¦

ā€¦ Iā€™m not sure A+ ā€œSysOptimizerā€ actually is the reason why
A+ also has two different ā€œInteger Modesā€ that do sound different, with Mode 2 being closer to Roon (still a very tiny little bit less ā€œdark-ishā€ and with ā€œan hintā€ more finest detail retrieval) and Mode 1, albeit being too bright on some systems, sounding definitely better than Roon (latest A+ version)

This is what I ā€œfeelā€ using USB out from the machine running Roon Server, into my headphones system
Canā€™t compare A+ and Roon on my main, speaker, system as Iā€™m using an Aries as Roon output, there.
To tell it all, though, I feel the Aries sounds ā€œa littleā€ better on its own than when used as Roon output

But in the endā€¦ Roonā€™s UI is so gorgeous and it does indeed sound so good that I donā€™t see any point in using something else anymore :slight_smile:

39

The only time I use Roon on my Core PC is listening to news via Roon Internet Radio stations.
That sounds ok.
I can understand the speaker.:sunglasses:

40

This is sound and well reasoned advice. You speak to the virtues of A+ for those using Mac-based systems. For those with Windows-based systems, HQPlayer is the counterpart (although HQP also plays on Macs). I know that Esoteric has done a huge amount of work on its own internal upsampling process and that may be a significant part of your great sound. But for many, if not most, other DACs the ability to upsample to high rate PCM (for Chord DACs and some others) to DSD 256 or 512 (for most other DACS) creates the possibility of using better filtering tools than are used in the DACs themselves. Roon now offers that capability in Roon, but most HQPlayer users feel that the different filter choices afforded in that program best allow us to fine tune the sound of our DACs to their highest level. Like Roon, HQPlayer also seems to benefit from putting its core on a very high powered computer and then using an NAA (network audio appliance) as the simpler, cleaner device connected to the DAC. Because of that neither Roon nor HQPlayer do what A+ does in optimizing computer system processes to reduce noise. There are, however, Windows optimizer solutions that some claim do improve sound by reducing system noise even in the main computer.

The other confusion worth addressing is that people confuse upsampling with hi-res original files and assume that the benefits of both can only occur at frequencies near or above 20kHz. That misses the point. Upsampling doesnā€™t recreate data that wasnā€™t in the original (say 16/44) file. What it does do is to allow the impact that the D-to-A conversion process has on the original to be moved further away from the audible range (so that any artifacts of that process are well beyond 20kHz) AND it allows the use of filters that introduce less pre-and-post ringing and less destruction of time-correct data than if those filters are applied to a 16/44 file directly. If poor filtering was used in creating the 16/44 file in the first place then much of that damage may already have been done (and that is why some of us buy hi-res original files because they at least imply care in handling in this regard). So the reason that upsampling matters in the audible range between 20Hz and 20kHz is that it prevents (as much as possible) artifacts of the D-to-A conversion process from spilling down into those frequencies.

6 Likes