NAS versus PC - any sound quality difference?

[LarsHP]

I am considering a NAS instead of using my PC since I would like to be able to play music without having my laptop/PC turned on.

At one forum I read someone saying that the sound from a PC/Mac based system was better (less jitter, I expect) than when running from a NAS or USB HDD to the SB Touch.
From my point of view a difference (if any) regarding jitter using a NAS vs. a PC would be that the CPU speed and perhaps RAM speed of the PC is higher and thus timing of the data stream could be more precise.
However, I am no computer engineer and there could be many factors that make this much more complex than just this.

I would like those who have experience with both types of setup to chime in on this - especially if you have high end audio gear.
I would also like those with deeper technical expertise to comment on any theoretical/technical explanations for any (or no) difference.

My chain in the digital domain is as follows at the moment:
FLAC files on an external HDD connected through eSATA to my (Windows) laptop running Logitech Media Server - ethernet - SB Touch (Fidelity Audio digital hardware mod) - Audio-gd Ref 7.1 DAC.
I would like to change this to: Files on a NAS running LMS - ethernet - SBT - DAC.

Above all:
Please don't start a general discussion regarding jitter. Let's assume, that jitter does play a role in digital audio.
However, I am open to experiences of no difference between using a NAS and a PC. In fact, I actually hope that there is no difference.

[alfista]

There are no relevant technical differences when it comes to delivering data to the SB between PC/Mac and NAS, some varieties of NAS are architecturally nearly identical to your average desktop PC. As long as it is powerful enough to deliver data at a fast enough rate there will be no difference.

[garym]

There is no difference.

Understand that the entire point of the TOUCH is that there is NOTHING about the audio/sound card aspects of the computer or NAS that has anything to do with the audio reproduction of the TOUCH. Much of the discussion about sound quality and computers has been in the context of all the things happening inside the computer, the actions of the sound card (and whether resampling, etc. is going on with the card, etc.). The SqueezeBox approach is to avoid all these potential issues (it is why the original creators of the SB line created these networked music players!). The computer serving the TOUCH doesn't even need a soundcard. It has nothing to do with the TOUCH. All that matters is that the DATA is sent to the buffer of the TOUCH. So the key is that you have a robust network. And if you use a NAS, it needs to be a high enough power NAS to run the program LMS (to handle browsing the music database, serving the data to the TOUCH, etc.).

[aubuti]

Let's assume, that jitter does play a role in digital audio.

Well it definitely does not play a role in the transmission of the data via ethernet (or wifi) between the pc/NAS and the SB. TCP/IP has no "timing" beyond making sure that enough data gets to the SB on time to avoid dropouts (and keeping in mind the Touch's reasonably large buffer). If you want to assume that jitter plays a role in TCP/IP then you won't have a fact-based discussion.

[Soulkeeper]

NAS versus PC - any sound quality difference?

Short answer: No.

A little longer answer: The hype about jitter having any effect in this context is just that. Hype. Squeezeboxes are asynchroneous wrt Ethernet. In fact, Ethernet is asynchroneous in its nature. So don't believe the magical thinking and the noise machines who insist on it.

thus timing of the data stream could be more precise

Ethernet is asynchroneous. There is no data stream, at least not in the way your sentence implies. All the data is transmitted in packets, and as long as all the packets reach their destination, the order or timing of the packets is irrelevant. As in, mathematically/logically irrelevant. In other words, completely, utterly, totally irrelevant.

The whole point of the Squeezebox is to turn those discrete packets back into a "stream" of audio (reassemble them, so to speak). The Squeezebox is not a sound card. It is more comparable to a whole computer with a soundcard. As long as the 0's are zero and the 1's are one, everything that has to do with jitter and timing, is the responsibility of the Squeezebox and the Squeezebox alone.

Just like, when your friend sends you a FLAC file as an email attachment, the qualities of the computer he sends the FLAC file from, are irrelevant to the sound quality that you will hear when you play the FLAC on your computer. This is a pretty direct analogy to what happens when your server or NAS sends a FLAC file to the Squeezebox.

If you want to learn more about this, I suggest you start with Ethernet. Wikipedia has plenty.

[Mnyb]

Do a forum search this has been debated to no end the forum is full of them .

But the gist of it is as long as the network can fill the players buffer it does not matter it is totally asynchronous .
That is in fact one of the main points with having a squeezebox it's sound quality is totally decoupled from the server .

The Touch buffers up to >30 seconds of a 16/44 flac so you can test it with no network by simply pulling the plug .
And if you are at the end of a playlist the buffer IS full and nothing more is sent from the server .
I never heard the last 40 seconds of my files sounding radically better ?

Jitter is not a real issue in the network interface.
And is often used as a catch all phrase for just about anything mysterius in the audiophile folklore ?

Most sources may have jitter on the spdiff and Toslink but very rarely on levels that are close to what's considered audible in controlled tests .
especially modern components .
Spdiff may be a conduit for electrical noise in some setups .
Spdiff is theoretically better but in practice it can be the Tos links that works best in some setups .
I'm using the spdiff out to the equipment you see in my signature .

I have used A desktop PC pentium4 dual core Win-XP and Ubuntu , A single core via 7 mini itx ClarkConnect linux and Now a HP microserver 36L dual core amd low power cpu with ClearOS linux .
I'm using the spdiff out to the equipment you see in my signature .

However the faster machine will give a much better experience with user interfaces Pad apps and foremost transcoding .

When you shop for a NAS get one with x86 architechture, so thaat SoX and FLAC and LAME will work fast without much strain for the CPU .
For example transcoding the ocasional 24/192 file to a Touch or the more common 24/96 files for use with a boom or duet .

One of the best tweaks is to not have the PC/NAS close to the hifi system as none of those things are silent (what ever the manufacturer claims ) .

Do not use screened ethernet cables as these are a way for noise transmission and ground loops in many cases .

Do read some of the old treads .
People that are far more knowledgeable than me have gone to great lengths to explain why the server won't matter a tiny bit re sq

For enlightenment and fun read this old tread .

http://forums.slimdevices.com/showth...-Optimisations

[garym]

Well it definitely does not play a role in the transmission of the data via ethernet (or wifi) between the pc/NAS and the SB. TCP/IP has no "timing" beyond making sure that enough data gets to the SB on time to avoid dropouts (and keeping in mind the Touch's reasonably large buffer). If you want to assume that jitter plays a role in TCP/IP then you won't have a fact-based discussion.

From my point of view a difference (if any) regarding jitter using a NAS vs. a PC would be that the CPU speed and perhaps RAM speed of the PC is higher and thus timing of the data stream could be more precise.

As aubuti correctly points out, there is no aspect of a datastream being "precise" when it comes to TCP/IP getting the data from the computer/NAS to the TOUCH via eithernet or WIFI. It either arrives or doesn't. If it doesn't, there won't be subtle sound quality differences; rather there will be dropouts, buffering, etc. Be careful when reading internet audiophile forums. Much of the discussion is entirely wrong when it comes to how a squeezebox system works. To be fair, the incorrect info is in many cases a "carryover" effect from folks assuming the old problems of "soundcards inside a computer" are still applicable to a SB player. And as I hope you understand now, this is simply NOT the case.

p.s. I started with a ReadyNAS duo for my SB server and it worked, but it was painfully underpowered, so things like browsing my music collection (about 70,000 tracks) was painfully slow. Scanning the music was painfully slow (all day versus 30-40 minutes). My vortexbox servers (which are simply low-powered computers without keyboards or monitor) are extremely fast on all accounts.

[LarsHP]

Well it definitely does not play a role in the transmission of the data via ethernet (or wifi) between the pc/NAS and the SB. TCP/IP has no "timing" beyond making sure that enough data gets to the SB on time to avoid dropouts (and keeping in mind the Touch's reasonably large buffer). If you want to assume that jitter plays a role in TCP/IP then you won't have a fact-based discussion.

As I said - I am no computer engineer, but my statement was regarding timing in the digital domain in general, not certain parts of the digital domain.

Thanks to both of you guys for your quick replies.

Do you think a ReadyNAS Ultra 2 RNDU2000 will do a good job?
As I understand, the NAS should preferably be x86 compatible ...

[Mnyb]

As aubuti correctly points out, there is no aspect of a datastream being "precise" when it comes to TCP/IP getting the data from the computer/NAS to the TOUCH via eithernet or WIFI. It either arrives or doesn't. If it doesn't, there won't be subtle sound quality differences; rather there will be dropouts, buffering, etc. Be careful when reading internet audiophile forums. Much of the discussion is entirely wrong when it comes to how a squeezebox system works. To be fair, the incorrect info is in many cases a "carryover" effect from folks assuming the old problems of "soundcards inside a computer" are still applicable to a SB player. And as I hope you understand now, this is simply NOT the case.

p.s. I started with a ReadyNAS duo for my SB server and it worked, but it was painfully underpowered, so things like browsing my music collection (about 70,000 tracks) was painfully slow. Scanning the music was painfully slow (all day versus 30-40 minutes). My vortexbox servers (which are simply low-powered computers without keyboards or monitor) are extremely fast on all accounts.

...and much of the whole debate of digital audio is a carry over from the analog era and don't apply either and is largely made up of misunderstanding basic concepts .

And treat any "information" from TAS or computeraudiophile or similar cult furoms as nonsense

[garym]

Do you think a ReadyNAS Ultra 2 RNDU2000 will do a good job?
As I understand, the NAS should preferably be x86 compatible ...

There are users here that use the readynas ultra and report excellent results. My understanding is that the readynas ultra is an x86 machine (which is important).