[DBB1]

Interesting. It never occurred to me that Linux might sound better than Windows. Is the Linux interface pretty much the same as Windows? Can you tell me more about your setup? What version of Linux? What hardware?

 

[GioF71]

Hello,.

  Interesting. It never occurred to me that Linux might sound better than Windows. Is the Linux interface pretty much the same as Windows? Can you tell me more about your setup? What version of Linux? What hardware?

 
Sure.
I do not really think something so definitive as saying "linux sounds better than windows".
What I think and have observed is that there is benefit when the machine you use to play your audio files does nothing else but audio playback itself.
Also, I did not build a dedicated windows machine for a direct comparison, but considering the much higher hardware requirements to run recent Windows releases, I am not interested in such a build.
 
If you think about it, software like fidelizer and jplay (which run on Windows) just do that: they lower the priority of other processes if not terminate them, to make sure audio playback has the highest priority (maybe these are too few words to describe those apps, but please correct if I am not precise enough).
 
So, provided that you have the chance and the resources and the estate on your desk for a dedicated box instead, this approach does not really make a lot of sense to me.
 
Windows is a desktop o.s. that 'needs' to be fined tuned for audio optimization. OTOH, on a Linux installation, I would not use a Desktop Environment either as if would lead pretty much to the same issues: higher CPU/Memory usage for activities not directly related to audio.
 
So my Setup:
 
Supermicro X7SPA-H mainboard with Intel Atom D510 1.66GHz dual core HT
Linear PSU from italian brand ZetaGi (set @12V)
1x1GB + 1x2GB RAM
Currently running Ubuntu Server 15.04, 64bit, Low Latency Kernel
MPD (Music Player Daemon - http://www.musicpd.org)
UPMPDCli (Upnp Renderer plugin for MPD - http://www.lesbonscomptes.com/upmpdcli/upmpdcli.html)
 
I use Minimserver on another linux box (MinimServer runs also on several prebuilt NAS servers) to handle the library. It is very customisable and almost unique if you love classical music, allowing custom tags, etc etc.
 
On top of UpMPDCli, I use BubbleUpnp Server to render the Upnp Renderer compatible with OpenHome.
This is not strictly necessary, but allows to use programs like Linn Kazoo, with native support for Tidal streaming (lossless also, of course).
The older (and END-OF-LIFE) Linn Kinsky works too.
 
Linn Kazoo runs on Windows so you can control you music server from your Windows desktop.
No worries of skips, pops, hiccups when you load your Windows desktop with some heavyweight app!
 
Or, if you prefer, you can control the box using BubbleUpnp on Android, which also supports Tidal streaming.
There are other Upnp controller applications availble, also for the Apple platforms. Here I am mentioning only what I have tried directly.
 
Also with Minimserver, using DSD transcoding to DOP, I am able to play DSD with application which do not natively support it. One is Kazoo which does not support DSD natively, as Linn does not believe nor support DSD as you may know.
 
If you are interested, I can provide more details.
 
I am not the white fly here, look at this new piece of hardware from exaSound, the PlayPoint Network Audio Player:
 
http://www.exasound.com/PP/Overview.aspx
 
It is more or less the same thing. Of course, the power supply of the exasound unit will probably be somewhat better than mine, no doubt on this. Also the initial configuration will be easier.
And it will probably sound better than my box.
But for a much higher price.
Software wise, I am basically using the very same stuff. Also, in case you notice, the HQPlayer NAA can be install on the linux platform. I also tried it, but I am sticking with the standard mpd + upmpdcli + BubbleUpnpServer at the moment as I did not notice any relevant advantage.
There are also solutions on linux for Airplay, but I have not considered them as I do not need this feature at all.
 
Keep in mind that all the software I used is Open Source or free (Linn Kazoo/Kinsky).
BubbleUpnp instead costs a few buck if you want to get rid of the ads (it's worth the very little expense).
 
Hope this helps!
 
Cheers
Giovanni

 

[DBB1]

Do you need a a Linux driver to run the U12?

 

[GioF71]

Do you need a a Linux driver to run the U12?

No, both the u12 and the x12 are natively supported.
And iirc the same is true also for the x20.

 

[Triplefun]

This is an interesting 'modern' perspective on audio quality and OS'
 
http://archimago.blogspot.co.nz/2015/08/measurements-audiophile-sound-and.html
 
.. The bottom line is that I am unable to find evidence to say that OS changes will make any difference to analogue audio output from a modern asynchronous USB DAC despite having heard comments already that somehow upgrading to Windows 10 will make things sound "better". Likewise, I see no evidence that software like Fidelizer will do anything to the sound output even though I can show that it "worked" to reduce the number of processes and threads of execution in the OS. Finally, and yet again, JPLAY demonstrates no clear ability to affect the sonic output and in fact (again after 2 years since the last test) I see problems with Kernel Streaming in playing back 24/48 accurately. This is obviously worse than "no effect" in that it actually deteriorates the quality of playback. I seriously wonder how these developers actually perform testing on their software and what they're actually doing! (Beyond marketing and dreaming up terminology like "ULTRAstream" for example.)
 
 
 

 

[Marcin_gps]

 
This is an interesting 'modern' perspective on audio quality and OS'
 
http://archimago.blogspot.co.nz/2015/08/measurements-audiophile-sound-and.html
 
.. The bottom line is that I am unable to find evidence to say that OS changes will make any difference to analogue audio output from a modern asynchronous USB DAC despite having heard comments already that somehow upgrading to Windows 10 will make things sound "better". Likewise, I see no evidence that software like Fidelizer will do anything to the sound output even though I can show that it "worked" to reduce the number of processes and threads of execution in the OS. Finally, and yet again, JPLAY demonstrates no clear ability to affect the sonic output and in fact (again after 2 years since the last test) I see problems with Kernel Streaming in playing back 24/48 accurately. This is obviously worse than "no effect" in that it actually deteriorates the quality of playback. I seriously wonder how these developers actually perform testing on their software and what they're actually doing! (Beyond marketing and dreaming up terminology like "ULTRAstream" for example.)
 
 
 

FYI: JPLAYDriver uses 32 bit int ASIO format for data exchange with host player (foobar in his example). (We chose that particular format as we believe things like data alignment are important for performance and other reasons).
 
That means driver _always_ gets 32 bits from foobar (so even for e.g. CD format) but has no knowledge of how many bits are actually used so it uses a simple rule: for 44kHz & 48kHz material 16 bits are used and any higher frequency is assumed to be 24 bits. (BTW 48kHz also used 24 initially but we went back to 16 after receiving user feedback from owners of older DACs (pre-HiRez wave) which support only 16 bits which made playback of 48kHz material impossible if 24 were used)
 
To play back 48/24 (btw, also 44/24) via JPLAYDriver, instead of 'Native' Bitstream needs to be set to '24' (if DAC supports it) - it's that simple.
It's not a perfect solution but one that works for most people. (Most HiRez is 96kHz & up and 48kHz/24 material is comparatively rare)
 
We actually recommend JPLAYmini or JPLAYStreamer be used for optimal results (as then we can eliminate influence of host player totally) and it just happens they have full knowledge of track format as they perform track loading themselves (which is technically not possible when using ASIO) so for any format Bitstream can stay 'Native'.
 
Cheers, 
Marcin

 

[scontin]

Last question and then I'll buy.
 
Better Gustard X20U or Audio-gd Master-7?
 
Double Sabre or R2R 1704UK?
 
Reading all thread, it seems that Gustard's boys have done the miracle to erase the sharpness and glare of Sabre chips.
 
In other threads I read Sabre signature is ever the same, though with some improvements using quality components.
 
@stuartmc have spoken about R2R.... other opinions are there?

 

[conquerator2]

The DAC19 would be a more fair comparison.
Hard to choose, unless you know exactly what you want

 

[scontin]

Hard to choose, unless you know exactly what you want

 
Warm and natural sound......

 

[conquerator2]

Warm and natural sound......

Well then I give the nod to the DAC19. IMO.YMMV.

 

[SimonBromley]

Hi there
 
I'm very interested in this particular dac. The feature set and the price for the performance on offer make it very tempting.
 
I've recently set my big rig up again after a long time in storage, and find the combination of Copland CDA 822 cd player ,
Pathos pre/power amp and huge old heavily customized Sony APM speakers, a bit too polite and easy going for me now.
Therefore, I'm after something to perk the system up and give it a bit more energy and dynamism. As well as drag it 
into the modern computer age.
 
The X20 sounds ideal, with out breaking the bank !! My only concern is that top end 'glare' I'm hearing about. A slight 
excess of high quality treble I can deal with via crossover mods if necessary. However, anything too shrill or piercing
would be a no go for me.
 
My Copland has an AD1853 converter, which I believe is a multibit Sigma/Delta device, and is very smooth and natural 
sounding. I've never winced at the treble from it. What I'm after is a similar-ish sonic signature, but a touch more dynamic.
 
Therefore, any further reviews of the overall sound from those who have this dac would be very much appreciated.
 
I know that purchasing this item will be somewhat of a leap into the dark regardless of the posts here, but I would like 
just a little bit more confirmation and description to settle my mind before I finally hit that purchase button.
 
Many thanks in advance.
 
Simon

 

[Triplefun]

Hi Simon
 
I upgraded to an avm pa5.2 hybrid tube preamp which has a 'warm'  sound similar to the pathos and found it really tamed the x20u without losing transparency and dynamics. I also use an old arcam dv88+ CD with Wolfson DACs and much prefer the x20u.

 

[GioF71]

 
This is an interesting 'modern' perspective on audio quality and OS'
 
http://archimago.blogspot.co.nz/2015/08/measurements-audiophile-sound-and.html
 
.. The bottom line is that I am unable to find evidence to say that OS changes will make any difference to analogue audio output from a modern asynchronous USB DAC despite having heard comments already that somehow upgrading to Windows 10 will make things sound "better". Likewise, I see no evidence that software like Fidelizer will do anything to the sound output even though I can show that it "worked" to reduce the number of processes and threads of execution in the OS. Finally, and yet again, JPLAY demonstrates no clear ability to affect the sonic output and in fact (again after 2 years since the last test) I see problems with Kernel Streaming in playing back 24/48 accurately. This is obviously worse than "no effect" in that it actually deteriorates the quality of playback. I seriously wonder how these developers actually perform testing on their software and what they're actually doing! (Beyond marketing and dreaming up terminology like "ULTRAstream" for example.)
 
 
 

 
So they are all the same in your opinion?

 

[GioF71]

   
So they are all the same in your opinion?

 
Also: in that article, a E-MU 0404USB is used to measure the performance of a TEAC UD-501 DAC.
 
From the specs of the E-MU 0404 USB:
Premium 24-bit/192kHz A/D and D/A converters* (A/D: 113dB SNR, D/A: 117dB SNR) deliver unmatched USB audio fidelity
 
So seen that this ADC is used as an instrument to measure the UD-501, the results are not surprising, are they?
 
I am not defending JPlay or Fidelizer but I think more precise measurements are required to just say that it's all s**t.

 

[Triplefun]

I am not really in a position to comment since I have only ever used Windows. However, given the speed of todays processors and memory, advances made in the various OS, the quality of the USB interface, and the maturity of the ASIO and WASAPI drivers then I don't think the underlying OS will have much of an influence on the sound. My lounge system has a dedicated Windows 10 music player and rarely skips a beat.  I find upsampling Foobar to DSD128 provides the most satisfying sound experience.
 
I also use Windows 10 on my Office system (which is also the Music server) with optional ASUS DS7.1 sound card and Schiit Loki DAC. I have loaded the ASIO4ALL driver to support Kernel Streaming to the sound card and the Schiit ASIO DSD64 driver for the Loki. I much prefer the Loki output compared to the sound card. When I first tested the X20u I did not notice much improvement over the Loki which I put down to the quality of the preamp and speakers. I don't use headphones. Also, I do a lot more processing on my office system (i5 2500k clocked to 4.2GHz and running web browsing, word processing, spreadsheets, ripping) and rarely does it interfere with the music normally playing in the background.
 
My lounge system with the dedicated notebook music processor (JRiver benchmark score of 2200), X20u DAC, AVM PA5.2 pre amp, AVM MA3.2 power amp and magneplanar 1.7 speakers, is much more exciting than my office system, provides a stable well defined image,  and justifies the investment in the X20u. I have made minimal changes to Windows and have not gone down the path of deleting redundant services. The CPU load (i7 620m) rarely exceeds 10%. I did not notice any changes moving from Windows 8.1 to Windows 10. And I prefer Foobar to JRiver - I feel I have more control using Foobar and I like the tablet client interface. Also I don't have to pay an annual JRiver licensing fee.