[davidross]

Some argue that Windows is the worst sounding of the major OSs. Is that one of the reasons that the Sonore servers are Linux-based or might that point not be relevant for these special purpose devices? I ask this in part because for me (and probably many others), dealing with a Windows-based device would be much easier because of familiarity, though if there are strong performance reasons for using Linux, they'd take precedence.

[skunark]

With the same hardware and playback scenario (local file, streaming, etc), any OS has the equal potential to sound the best.    At the end of the day, it's about getting the PCM audio data to the buffer used by the soundcard/device.   What makes Windows more challenging is how the audio application communicates to the sound card/device through a virtual driver.  In other words, the audio playback application has to go through a middle man (universal audio device) when using MS approved drivers.  What's even more challenging is how other applications can impact the middle man, just by starting an application or updating the version can cause custom settings to change.   

Windows 8 has introduced an audio hardware acceleration path which enables custom drivers or even the application to gain exclusive access to the hardware, but has limited hardware support. 

With that said, Mac OS and Linux can have all this same exact issues, just with linux, you become your own boss on how to configure and update the OS and applications and if you don't like the options available, you can change the configuration or write a little C/C++ to create it.   

At the end of the day, you should be able to configure Windows, Mac OS and Linux to all be bit-perfect, but in a consumer product, i would want complete control when it comes down to reliability of playback.

Jim

[davidross]

Thank you, Jim - that's very helpful. I wonder if it's safe to assume that Chris Connaker's Windows-based servers provide bit-perfect transfer. I should write to him about this, too, before making any purchases.

[barrows]

David: just a quick follow up.  Sure, windows systems can be made to work for audio, but it aint easy to get the best out of windows, hence things like jPlay+jRiver+audiophile optimizer which some people are using.
Sonore serving devices use linux based OS, because linux can be configured to only have what one needs to play audio, and nothing else, from the ground up.  This is why you see pretty much all high end commercial music server products using linux variants (like Aurender, etc).
Sonore devices are plug and play, all the configuration has been done already, with prorietary linux based OS configured just for playing music, and nothing else.  These devices are very easy to use, and require no intervention from the owner in terms of configuration.

[catastrofe]

I have been a long-time Sonore customer after spending several years with Mac based solutions.  IME, Linux just sounds better.

[firedog]

I went from Windows Home Server to Vortexbox to Windows 7 on the same server. Once I had it setup I noticed no SQ differences between the OSes. Windows 7 and 8 have no problem playing bit perfect as long as you use an ASIO, Kernel Streaming, or WASAPI driver. Not exactly difficult to setup. If you use a Windows server as an audio only server it doesn't have many processes running and isn't inferior to the other OSes. There are still a fair number of audio devices that aren't Linux compatible. Hopefully that will change, but not all manufacturers see being Linux compatible as a necessary feature.

One of the additional reasons bespoke server manufacturers use Linux is that it is free. It also runs better on low power devices and many audiophile PCs/streamers are fairly wimpy in computing power and memory. It's much easier to keep costs down on small devices if you run less expensive HW and don't pay for expensive high power CPUs and lots of memory. It's easier to build a fanless server with Linux b/c you can use lower power devices.

I have nothing against Linux, but IMO it is an audiophile myth that it is inherently better sounding than Windows.

In my particular case I found myself intervening more with Vortexbox/Linux than I do with Windows. My specific MB had issues with VB. Worked, but not without some help.

One of the reasons that Chris uses Windows on his CAPS servers is that Windows gives the user a much larger choice of software to install and use on the server for various audio needs, and for the average user Windows is much easier to use than Linux. I know this isn't true for people who are familiar with and like Linux, but for about 95% of the listening public it is true.

I'm really not trying to debate Windows vs. Linux. I have zero argument with someone who prefers Linux, for whatever reason. If your experience is that it does sound better, then go with what you know. Just trying to bring a different perspective. On audiophile sites the users are much more sophisticated than the  the general - or even the moderately sophisticated - public in terms of using and installing software.

From Computer Audiophile:

All four CAPS servers will run Windows 8.1 Professional 64-bit and JRiver Media Center v20. At this point in time Windows 8.1 is the only stable game in town that's compatible with any number of applications and audio components. I'm excited to see what members of the CA Community will do when they install Linux, but for now my four main designs are Windows based. I selected JRiver Media Center v20 because it's very stable, has great support, is widely used in the CA Community, and is capable of great sound quality for all PCM and DSD sample rates as well as multi-channel audio. There are other applications available that I consider a bit more fringe at this moment in time. I'll leave it to the CA Community to use and tweak those applications and share results with everyone. As a formula and platform I believe Windows 8.1 Professional and JRiver Media Center v20 is the best combination.

[shadowlight]

This is why you see pretty much all high end commercial music server products using linux variants (like Aurender, etc).

It could also be due to Linux being free to roll v/s Windows which requires a license.

[catastrofe]

I think you will find that go get a Windows or Mac based system to sound as good as Linux, it requires a lot of optimization.  This has been my experience over 6 years of computer audio involvement.  I've opted for the simpler route, but there is certainly more than one way to achieve excellent playback.

[barrows]

I went from Windows Home Server to Vortexbox to Windows 7 on the same server. Once I had it setup I noticed no SQ differences between the OSes. Windows 7 and 8 have no problem playing bit perfect as long as you use an ASIO, Kernel Streaming, or WASAPI driver. Not exactly difficult to setup. If you use a Windows server as an audio only server it doesn't have many processes running and isn't inferior to the other OSes. There are still a fair number of audio devices that aren't Linux compatible. Hopefully that will change, but not all manufacturers see being Linux compatible as a necessary feature.

One of the additional reasons bespoke server manufacturers use Linux is that it is free. It also runs better on low power devices and many audiophile PCs/streamers are fairly wimpy in computing power and memory. It's much easier to keep costs down on small devices if you run less expensive HW and don't pay for expensive high power CPUs and lots of memory. It's easier to build a fanless server with Linux b/c you can use lower power devices.

I have nothing against Linux, but IMO it is an audiophile myth that it is inherently better sounding than Windows.

In my particular case I found myself intervening more with Vortexbox/Linux than I do with Windows. My specific MB had issues with VB. Worked, but not without some help.

One of the reasons that Chris uses Windows on his CAPS servers is that Windows gives the user a much larger choice of software to install and use on the server for various audio needs, and for the average user Windows is much easier to use than Linux. I know this isn't true for people who are familiar with and like Linux, but for about 95% of the listening public it is true.

I'm really not trying to debate Windows vs. Linux. I have zero argument with someone who prefers Linux, for whatever reason. If your experience is that it does sound better, then go with what you know. Just trying to bring a different perspective. On audiophile sites the users are much more sophisticated than the  the general - or even the moderately sophisticated - public in terms of using and installing software.

From Computer Audiophile:

firedog:  The original question was regarding specifically, Sonore server devices, and your response mostly pertains to DIY solutions.  With Sonore devices there is no set up or user intervention required as far as the OS goes, they are plug and play, so your concerns about set up do not apply.  If a Sonore customer does have an issue (which is rare, my custom server has been running now for two years without any issue, just start it up and it works) Sonore also provides the best support in the business, where Jesus will fix any issue which shows up.

As for campatibility issues with external devies, I assume you mean USB DACs, right?  Sonore devices, and Linux in general, is compatible wiht the USB Class II audio standard, so any dAC which complies with the standard is good to go.  The problem is thta a few manufacturers (and this is less and less every day) do not make their USB receivers compatible wiht the USB Class II audio standard, these non compliant DACs always have challenges, as the manufacturer usually has to develop a new USB driver every time there is a version change of the OS-these challenges do not apply to DACs which comply with the USB class II audio standard (except with dumb Windows devices, for some reason Microsoft will not make Windows compliant with the USB Class II Audio standard either, annoying).

Additionally: it is precisely because Linux allows for a very low power computer which provides a sonic benefit.  Higher power processors, bigger power supplies, more RAM, etc, etc, all result in more noise from the computer with the potential to harm audio playback performance.

[davidross]

Thank you all for this very useful discussion. Regarding power issues, I've wondered whether the Vortexbox appliances from Small Green Computer are more likely than those from Sonore to introduce noise because media storage and CD ripping hardware are internal to the enclosure. There may also be an inherent disadvantage in using a single drive for the system and library which is the most straightforward way to use the SGC boxes. Does all that seem right to you?