ND5 XS: Different NAS, different sound

No, its more likey you believed you heard a difference, even though there cannot be one.

Garyi

You sound like a doctor "No you're not ill, because I don't  understand what you have".

Dear Garyi,

Thanks for sharing your view.

I am with you and I sincerely hope that this is the case.  Life will be so simple then.  Unfortunately, the fact is that there is indeed audible difference when using the new (Synology 712) and the old (Synology 209) NAS to stream the music across to my beloved ND5 XS.  712 excels in terms of the sound stage, the micro-dynamics, contrast, tonal balance and overall presentation.

I am trying to come up with some logical explanations for this finding.  The only "plausible" (but admittedly not scientific) explanations are:-

1. The transformer used in the new 712 model (6.8A) is of better quality than that of the old 209 model (5.8A).
2. The DSM version (i.e. operating system) used in the new model (version 3.2) is better than that of the old model (version 4.0). 
3. The old Synology has not been used for around 2 month.  It takes a bit of time to warm up.
4. The processing power and amount of RAM do make a difference.

I know...I know.  All the above lack the support from 1010 world. Both the new NAS and the old NAS has enough throughput for the purposes of streaming the music.  In theory, the raw music data gathered from the same hard disk and streamed through either the new or the old Synology model should be identical in every case.  In reality, there is a difference and I missed my 712.

Thete is one reason that can cause a difference and I have heard it with others on a very revealing system into an NDX/555PS NDAC/555PS and 5 series amplification. (on a blind listening test we correctly identified differences 100% of the time). This appeared down to the TCP windowing parameters. A NAS (and specially the uPNP server) thati uses small TCP windows puts more of a load on the peer TCP stack in the NDX streaming board. A uPNP server that lets the window size grows to the largest the NDX can happily handle seems to sound less 'dry' and attractive.

I conducted this experiment with the original streamer boards, may be different for the newer.

The only way to see what is happening here is to use a TCP/IP sniffer such as Wireshark.

Finally the difference is driven by the upnp server TCP stack. If the NAS is seperate from the upnp server then we have seen and heard no differences from different NASs 

Simon

I don't think you're alone. I too have a Synology Nas. It's the DS210+ and I think I'm hearing significant differences when I stream from it to my SU compared to when I stream from my laptop. On first impressions the NAS seems much worse but I'll have to listen over a longer period. Thanks!

There certainly could be major differences due to settings between the two.  If you had identical settings I would say, there is no way.

Perhaps your previous NAS/Server was transcoding to mp3 or something.

None of these would affect sound quality, unless the lower spec'd unit was experiencing drop-outs or something.

Make sure you have identical settings in both machines.

Simon,

You had mentioned this before. I appreciate your plausible explanation. However, I would not rule out the possibility of any other factors we may have not thought of before. Sonic differences btw different NAS has been reported a couple of times in this forum and in other forums. I would not be surprised if some "NAS test" are published soon. I don't know how such tests could be conducted because such influences will be very much network and  system dependent.

Martin Colloms of HiFi Heretic Critic has compared and written about audible differences between some digital storage devices. The full text has been posted on the Enjoy the Music website. Google " Listening to digital storage" and it should come up first, right under the Google ads.

Jan

Dear pcstockton,

Thanks for your reply.   I agree wholeheartedly with you that my propositions (e.g. better spec of the NAS yields better sound) defy logic in the context of audio streaming.  However, in reality, the sound generated is different even when the NAS in question is partnered with identical hardware.

Regarding the software settings (e.g. trans-coding in the DSM), all my music files are in AIFF format.  Naim supports AIFF format.  Accordingly, even if I have accidentally switched on the trans-coding function of the NAS, I believe that such function will not be performed.

On a related note, is there any difference when using a traditional spinning hard disk as compare to a solid state disk inside a NAS?  Adopting your logic, there should be again no difference in the sound generated through the streaming process. 1 is still 1 and 0 will still be 0.

bb7668,

What i mean is that one of the servers may be transcoding to mp3 or something.

-p

If some accidental transcoding is going on, you will be able to see this on the your Naim unit when you view the stream information, I know it's not happening in my case. Thanks!

Dear Nagoya,

As you have rightly mentioned, I always check the stream info window.

You're not alone. I tried the DS212J and thought it sounded bad. The DS212 (1.6 Ghz) sounded a lot better.

Compared to the DS212, with the DS212J, it sounded like the system was under some strain. So much so that when I used a streamer (SBtouch) into the super uniti through an optical connection it sounded a whole lot better.  Just not that impressed with the unit because of that reason. Probably better to separate out the analog and digital side through different power supplies.

I'm now getting a QNAP (T219 with 2Ghz processor and 512 MB Ram) for the same price as the DS212.

Aysil, you are right in there is always the  possibility of new not yet discovered phenomena, but the TCP windowing has adefinite cause and effect in sound quality on the NDX. TCP windowing and stack loading is clearly understood and is fundamental TCP characteristic. No voodoo or black cats... And upnp servers in NASes with limited memory on the TCP stack will NACK away like a sewing machine, where as a larger buffer stack will allow larger windows and require less NACKs and be more like a loom by comparison.

However I am interested to hear whether the more rigorous decoupling/ screening of the streaming boards on the NDS (assuming they handle the TCP stack in the Naim architecture) reduces prominence  of this artefact.

Simon

Dear MangoMonkey,

Thanks for sharing your experience.  I am glad that I am not alone. I particularly like your description of the sound (i.e. under some strain) when comparing the DS212 and its stripped down J version.  This succinctly summaries the "predicament" I am facing right now.

There is always a belief that, so long as the NAS is capable of channeling the raw music data to the streamer without dropout, the job is done.  Apparently this is not the case in the hifi world, at least with regard to my current setup.  Hifi is really a strange thing!

then spend $10K on a NAS

Dear pcstockton,

I am in no position to guess the tone of your latest comment.  In case you are remotely interested, there are indeed some more refined music storage systems/servers in the market.

http://www.aurender.com/

Simon

Very interesting. Can you suggest a site where I may find more info on TCP windowing.  I know I can Google it but your suggestion may get me to something more focussed.

Thanks

Here is a link to a NAS which costs more than $10K! It has 1Tb capacity, has a UPnP server incorporated, and has even a ripping engine included. It has a built-in dac section but a lot of people do not use this feature!

Sure, but where would be the fun in doing that? Much rather have a $500 system tweaked so it beats or even comes close to the $10K system.

On the other hand, it amuses me no end that people after paying for $10/foot Naca5, $875 for a powerline etc. decide that the must save the $200 and go for the cheapest Nas drive possible.

The whole selling point of DACs and streamers has been that they make the transport irrelevant.  Many apparently have bought into this wholesale and have the mentality of differences being impossible - everything that matters can and has been measured and corrected for - jitter, bit perfect files etc.  I guess they were wrong.  Quel surprise.  We've replaced one set of unknowns in digital replay with another.   

Makes sense to me.

1. A general-purpose computer can play CDs, but not as well as a dedicated computer built for that single function, i.e., a CD player.

2. A general-purpose computer can be used as a NAS, but a dedicated computer built for that single function  will perform better.

3. If I dare extend the analogy, a general-purpose NAS can serve music files over a home network, but not as well as a dedicated NAS built as a music server, i.e., the UnitiServe, or other similar appliance.

IMHO of course,

Jan    

PS. Since links seem to be OK here, here's the HiFi Critic review :