Sound quality issues with NAS

Roon Software Discussion
1

Hi to everyone.
I’ve been trying to find something related to my problem, however I need to start new topic.

I’m running Roon Core on my Mac Mini 2018, which is directly connected to Benchmarc DAC3 by USB.
The biggest part of my library is on QNAP NAS, which is in my local network. Everything is plugged by CAT6a cables into my ASUS AC1900 Router. LAN Cables are not long. From NAS to router approximately 1 meter, from router to Mac Mini about 4 meters.

I’ve been noticing from the begging that files played from Mac Mini’s local SSD sound much different, than those coming from NAS. Local files are brighter, more detailed, compared to those stored at NAS. Even comparing the same file stored at local disc or NAS, they sound clearly different.

What am I doing wrong? Is there something in Roon settings that I should pay attention to, or something in network settings.

Any help appreciated.

3

Network streaming is a can of worms in terms of sound quality. Everything in the signal path seems to degrade quality. Network switches make a difference to sound quality. Network cables make a difference. Devices such as LAN isolators make a difference. Power supplies for these devices make a difference.

The directly connected source (your Mac) becomes your benchmark and the goal is to find the network kit you need to match that (and it may not even be possible).

Switches from the likes of AQVox, SoTM, JCAT, Uptone, etc. Cables from same. LAN isolators from various sources, eg Pink Faun.

Can of worms but lots of discussion on hifi forums.

4

The end file being played is identical no matter where it comes from. If they sound different, it’s nothing to do with the file location.

Are the files identical? You don’t have copies accidentally ripped at different settings do you?

2 Likes
5

It isn’t audio until your DAC’s finished with it, until then its data, and we know the data is fine because it’s buffered and checksummed more than once on its path. I wonder if “expectation bias” is playing a role here? Have you tried making similar comparisons with the same file on a local USB drive?

3 Likes
6

And yet the OP hears a difference in sound quality. Explain…

7

Same challenge. In the “bits is bits” world, the OP is hearing a difference. Explain…

8

The files are identical.
Ripped by myself, using XLD for MAC.
I am running these A/B tests with the same files, stored on different locations.

9

Can you confirm the op’s observation, because this is certainly not my experience?

10

Mike, not his particular setup, but otherwise yes: in my system, streaming from a source connected directly to my front-end/DAC is superior to ethernet streaming through the same device (in fact, I’ve seen this on all network devices I’ve owned). I have, since first noticing this some time ago, substantially lessened the gap through experimenting with switches/cables/power supplies etc, but a gap still exists and so the work continues. Has caused a lot of head scratching as to the cause, for sure.

1 Like
11

Are you sure it is not the ethernet module in your DAC that is causing the trouble?
It is of course not the ethernet port itself, bit the 'integration module and corresponding firmware, that can have an impact on signal quality.
Also, can it be that on ethernet, you have a lower signal level than via USB. 1 dB difference is hard to detect, but allways result in lower perceived quality on longtime listening.

Personally, I do not perceive any difference when listening to either albums stored locally on SSD, stored locally on NAS, or directly from Qobuz via the internet.
I must admit however that the weakest link in my setup are my own ears (barely hearing up to 10 kHZ - if at all - mainly due to age).

Dirk

1 Like
13

“I don’t understand how or why there is a difference, therefore there CANNOT be one”
“I don’t understand how or why there is a difference, therefore there CANNOT be one”
“I don’t understand how or why there is a difference, therefore there CANNOT be one”
“I don’t understand how or why there is a difference, therefore there CANNOT be one”
“I don’t understand how or why there is a difference, therefore there CANNOT be one”
“I don’t understand how or why there is a difference, therefore there CANNOT be one”

4 Likes
14

Jesus Christ, not again.

Don’t you guys ever get tired of this useless back and forth?

5 Likes
15

OK. So you are completely dismissing his experience… How does that help him?

17

I’d suggest that it helps by suggesting OP looks elsewhere for the issue, as it’s not a file issue.

2 Likes
18

I see that some comments are rather sceptical.
Thanks everyone who tried to help.

I dismiss topic and my questions at this point.

19

Do you have the system connected to tidal or qobuz? Could it be choosing online versions in preference when you’re choosing to play certain files?

How do you tell roon which storage medium to play from, as in, do you literally rescan your entire library? Or do you keep two identical copies of the tracks named differently so you can A / B them back to back?

If you can tell us more about how you’re setting up the system to hear the differences, maybe someone can advise.

1 Like
20

Have you ever monitored the CPU usage on your Mac Mini when you listen to music, either SSD or NAS?

With Activity Monitor you can monitor it. Do you have higher CPU usage when playing the music from the NAS?

Does this higher CPU usage emit more RF on the USB cable?

Use once a USB Cable with a ferrit core between your Mac and the DAC.

There are cables from Delock and Lindy. The shipment for these cables is probably more expensive then the cables themselves.

If you want to go overboard get an Intona isolator. The industrial one. And use a USB cable with a ferrit core. That never hurts.

1 Like
21

Other question: How much RAM does your Mac have? How much is used, how much is free?

22

Neither of these would change depending on storage location though. They may be issues in themselves (although I’ve yet to find any link between free resources and sound quality).

If the system can play the file without skipping that should be sufficient to play it back perfectly.

The only reasons I can think of for differing quality are 1) the files being materially different
2) roon doing something weird and preferring an online version when one library is used
3) some differing audio setting when local storage is media source

23

You are right.

I started to debug. If looks to me like Roon reads the files into a buffer, no matter if it is local, NAS or coming from Tidal/Qobuz.

Then it applies the upsampling, DSP, EQ and send 5 seconds chunks to the RAAT client.

The RAAT client client maintains its buffer 100% full for the whole playing time. When the song ends, it empties the buffer.

I do not know how Roon actually buffers, the actual handling of the file and sending it to the client is the same, not mattter where it is originally stored.

Nothing which explains, why it could be different.