Jump to content
bluesman

Article: The Value Proposition In Computer Audio: Nuts, Bolts, and Building Blocks - Building a Home for Your Player Software

Rate this topic

Recommended Posts

A nice introduction to the simplicity of computers..... So much to take in but thank you... when my head stops spinning ill re read it. Amazing..

 

Dave

Share this post


Link to post
Share on other sites
7 minutes ago, Middy said:

A nice introduction to the simplicity of computers..... So much to take in but thank you... when my head stops spinning ill re read it. Amazing..

 

Dave

Hmmm - now that I see it on the web, it's looking a bit larger than it did on my monitor :) 

 

I could have split it in half, e.g. propriertary vs open source.  But I was so far into it by the time I realized what a monster I was creating that it was easier to just finish it, move on, and remember the lesson.  The next piece takes the same approach to single board / micro computers for audio, but is lighter and easier to digest.

Share this post


Link to post
Share on other sites

Wasn't a criticism at all, just what little i thought I knew, now needs an electron microscope to see it.  One aspect is convenience verse fear of trying. Windows is always install.exe and play. Linux frightened me a bit and gave up even AL and Daphile until i figured out Euphony enough for my skill level to set up with just the GUI.  The untapped knowledge and experience is coming out nicely with these articles. Really thank you Bluesman, yours and others efforts have really put AS into the next level with these articles...

Sincerely 

David

Share this post


Link to post
Share on other sites

Thank you for this amazing series of articles! 

 

Looking forward to the next one on SBCs. Just bought a RPi 4 to play with, mainly as a Roon bridge. I’ve been concerned that it might not provide what I’m hoping for, but they are so inexpensive why not give it a shot! 

 

Many, many thanks! 

 

John J


JJinPDX

Share this post


Link to post
Share on other sites

@bluesman

 

This is great and it will help a lot of folks out. 

 

I have found that the OS can matter but maybe less that I expected.  If the Playback software is "tuned" to the environment then the OS becomes less relevant.  I have not used Windows or macOS as the core software for my music playback system because I felt more obligated to manage it more tightly with OS updates and the like.  Several of the Linux distros that are "music" oriented have a web based console to control the operation.  To me this improves the overall usability of the dedicated computer for music.  Yes it limits in other ways.  Some of these distros run from a command line menu system using SSH.  This can be frustrating for some and way over the head of others.

 

I really have no skin in the OS selection game.  I do have some worries about the long term viability of say custom hardware with custom linux installed. I also worry about where Microsoft is going with Window 10 and their data collection efforts in Windows 10.

 

In the last couple of months I have been happily surprised by software on Raspberry Pi and on X86.  That is cool.

 

Thanks again.

 

Share this post


Link to post
Share on other sites

That was A LOT of content about operating systems. 

 

I'm sure there are many other ones and I assume other people will mention ones they figure you missed, but I want to throw Snakeoil OS (https://www.snakeoil-os.net/) out there as another 100% audio dedicated Linux distro. And it has Roon and HQPlayer configurations now too.

 

I've always thought the lead developer Keith was a great guy and very focused on being an audiophile and making an audiophile OS.

 

Share this post


Link to post
Share on other sites
9 hours ago, JJinPDX said:

Thank you for this amazing series of articles! 

 

Looking forward to the next one on SBCs. Just bought a RPi 4 to play with, mainly as a Roon bridge. I’ve been concerned that it might not provide what I’m hoping for, but they are so inexpensive why not give it a shot! 

 

Many, many thanks! 

 

John J

I think you’ll find it to be all you want and need. I run RB on a stock 4G Pi 4b in my living room system (SMSL SU8 into Prima Luna Prologue 5 driving Focal 726 towers) and it sounds fantastic! I haven’t rebooted in weeks, and the temp never gets above the low 40s C with a Flirc fanless case.  
 

I go into cases and cooling in a bit of detail in the next installment, but the one piece of info you need now is that a cool Pi will last for years in audio use - but if you let it overheat, it will probably die of a stroke in middle age.  You don’t need a fan to keep it comfortable, but you do need a good metal case like the Flircs that I prefer. I cite data on cases and cooling in the next piece, and the Flirc does the best of any passive case (& looks really nice).

 

You’ll probably never see it, but a thermometer icon will appear on your Pi’s GUI (if you’re using a video monitor) if the CPU temp hits 80C. At this point, it will throttle the CPU back to 3b speed to help cool it.

 

Thanks for the nice words! Enjoy your music.

Share this post


Link to post
Share on other sites
1 hour ago, Ben-M said:

That was A LOT of content about operating systems. 

 

I'm sure there are many other ones and I assume other people will mention ones they figure you missed, but I want to throw Snakeoil OS (https://www.snakeoil-os.net/) out there as another 100% audio dedicated Linux distro. And it has Roon and HQPlayer configurations now too.

 

I've always thought the lead developer Keith was a great guy and very focused on being an audiophile and making an audiophile OS.

 

Snakeoil is basically a bare bones Ubuntu server with a real time kernel. It leaves out the processes that aren’t needed for audio but that may affect it adversely (a controversial topic open to discussion elsewhere).  I’m sure it’s fine, based on extensive experience with many other Linux distros on dozens of boxes. But I find any benefits of a real time kernel for audio (playback only) to be somewhere between minimal and undetectable on good, modern hardware.  Multiple CPU cores and hyperthreading have largely eliminated these factors.

 

Even for recording, the only tangible benefit of a RTK over low latency versions (if there is any) comes when recording tracks alongside existing ones. Latency results in asynchrony between the monitored track(s) and the live one being added. This is compensated in good recording software, but the better choice is to eliminate it up front. There are adjustment settings in good DAWs to do this.

 

Theoretically, a real time kernel that prioritizes CPU calls in audio processing will reduce potential for any degradation of the bitstream, with allegedly better SQ resulting. I haven’t heard this myself, and I suspect that it would take a better system than I’ve ever had to demonstrate it if it’s real.

 

I do plan to try Snakeoil for RPi. It might benefit performance from the lower power versions (ZeroW and 3) and deserves a listen.

Share this post


Link to post
Share on other sites
2 hours ago, bluesman said:

Snakeoil is basically a bare bones Ubuntu server with a real time kernel. It leaves out the processes that aren’t needed for audio but that may affect it adversely (a controversial topic open to discussion elsewhere).  I’m sure it’s fine, based on extensive experience with many other Linux distros on dozens of boxes. But I find any benefits of a real time kernel for audio (playback only) to be somewhere between minimal and undetectable on good, modern hardware.  Multiple CPU cores and hyperthreading have largely eliminated these factors.

 

Even for recording, the only tangible benefit of a RTK over low latency versions (if there is any) comes when recording tracks alongside existing ones. Latency results in asynchrony between the monitored track(s) and the live one being added. This is compensated in good recording software, but the better choice is to eliminate it up front. There are adjustment settings in good DAWs to do this.

 

Theoretically, a real time kernel that prioritizes CPU calls in audio processing will reduce potential for any degradation of the bitstream, with allegedly better SQ resulting. I haven’t heard this myself, and I suspect that it would take a better system than I’ve ever had to demonstrate it if it’s real.

 

I do plan to try Snakeoil for RPi. It might benefit performance from the lower power versions (ZeroW and 3) and deserves a listen.

You covered it all pretty well in that post. It's worth a shot to try it out in a qualified listening environment, as in, not just with budget gear or something you don't care much for.  

 

I think it's a decent Linux OS for audio and more than qualified to play the part. But as you said, can I unequivocally say it's the best and beats all the others, no. I just think it's comparable and totally worth a spot in the list given what's being covered here. 

 

And I must say thanks for all the effort you put into this 😀

Share this post


Link to post
Share on other sites

Audiolinux, Euphony truly works. On the same HW it truly sounds way better than optimized Windows 10 with Roon. Music sounds very natural and musical. Noticeably better than allo digi one signature powered by good power supply. This is the first time I enjoy using PC as transport component.

Share this post


Link to post
Share on other sites
15 hours ago, bluesman said:

That’s great to hear, and your feelings are shared by many. My point about this example is only that there are dramatic interpretations on all sides of all such comparisons, and the superlatives used to describe each are pretty strong (e.g. “way better”).  The SQ gap between any two would have to be wide to go from “way better” on one side to “way better” on the other.
 

Given the close to equal opinion split between any two, I believe that these gaps are probably 90+% personal preference driven by confirmation bias, selection bias, etc and <10% real ordinal differences in SQ.  There are so many confounding factors that we can’t possibly control for them all, e.g. minor deviations from flat frequency response combined with nonlinear listener hearing sensitivity across the audible spectrum.

 

So I believe that listeners have definite and valid preferences and should use the systems they find most pleasing.  I just don’t believe that most of these differences are objectively as dramatic as the chosen descriptive terms suggest.

 

Well the last 3-10 % of SQ is usually the difference between enjoying the listening to music and not enjoying the music at all. And that is in the end the most significant difference...

Share this post


Link to post
Share on other sites

Yep - most definitely not style over computer.


We are far more united and have far more in common with each other than things that divide us.

-- Jo Cox

Share this post


Link to post
Share on other sites
1 hour ago, bluesman said:

You quoted me correctly but ignored what I said, which is that

 

"I believe that these gaps are probably 90+% personal preference driven by confirmation bias, selection bias, etc and <10% real ordinal differences in SQ."

 

I was only referring to the described differences in SQ between alternatives (the "gaps"), not SQ itself.  When one group of audiophiles describes the SQ of a component, system, etc of known high quality as "fabulous" and "heavenly" and another equally adept and experienced group says it's "horrible" and "unlistenable", there simply cannot be that much objective difference between the alternatives.  Had I chosen to shed all pretense of being PC, I'd have said that it was 50+% personal preference, 49+% semantic hyperbole, and SQ that's QNS.

 

If there were a system that truly reproduced and presented a recorded source exactly as it was performed, and if audiophiles were all equally able to both detect sonic differences and evaluate them totally without bias, almost everyone within a given budget would have the same system.  When a roughly equal number of audiophiles seem to prefer either of any two alternatives neither of which is perfect, the most logical conclusion is that the alternatives differ minimally and that listener-related factors are driving that decision rather than objective variance.

 

Parenthetically, I find it hard to understand how a music lover could "...not [enjoy] the music at all" because of SQ differences or deficiencies.  Most of us listen throughout the waking day on a variety of devices and systems that are suboptimal, e.g. smartphones, car systems, portable players, etc.  When SQ is marginal, most music lovers subconsciously apply "neural filters" that let us hear the music as it is despite sonic aberrations, so we can we can enjoy it even though reproduction is flawed.  Listen to a talented amateur string quartet play on student-quality instruments, or a concert in a venue sufficiently large &/or designed poorly enough to require sound reinforcement - the musical content is a joy despite its flawed presentation.  When you listen to the equipment, you hear the equipment.

 

Seems that  you’re man with issues. Not sure why they let you post your stuff here in the first place. This website is getting desperate I guess... 

Share this post


Link to post
Share on other sites
8 hours ago, mentt said:

 

Well the last 3-10 % of SQ is usually the difference between enjoying the listening to music and not enjoying the music at all. And that is in the end the most significant difference...

 

I agree with this ... when the SQ is 97% or so there, every annoying deficiency of the playback system keeps on nagging one; the 'signature' of the playback chain builds up as an ever present 'veil'; meaning that certain recordings can't be listened to with pleasure, because the characteristics of the recording and playback chain clash - the easy excuse of "bad recording!" is casually uttered, to explain the "inconvenient truth" away ...

 

Get the last 3% in place, and the transformation is, subjectively, huge - no longer a rig of equipment in front of you, trying its best; but a musical event, effortlessly presented.

 

Highly sub-optimal devices like car radios are fine, because one's expectations are completely different - you can enjoy watching kids mucking around on the tennis court; but when a highly ranked player is not at their best it can be quite frustrating to see them making one bad shot after the other.


Frank

 

http://artofaudioconjuring.blogspot.com/

 

 

Over and out.

.

 

Share this post


Link to post
Share on other sites

First, thanks for the very thorough review. With the enormous number of options even a review of this size must leave out a lot of information, but there is plenty enough here to be very useful.

 

I haven't used FreeBSD in a few years (on which MacOS is partly based), and wonder whether it has drivers for modern cards and DACs. There are a variety of free music players for it (including the old standby MPD). But it involves even more command line use than Linux. Great OS though, especially for server functions.

 

Interesting how different things bug different people. I've had no problems with wi-fi adapters with a variety of distros, but Samba configuration has occasionally turned out to be a pain for me.

 

Manjaro is a more newbie-friendly, GUI-centric version of Arch (it's what I'm on as I type this), and might be attractive to some folks. But if you are familiar with the Debian/Ubuntu/Mint way of doing things and just want to use your OS to get stuff done rather than embark on a learning curve, there's no reason to change.

 

I often use Xubuntu (Ubuntu with the XFCE desktop environment) because it was an easy distro to learn to use, and I've been accustomed to XFCE since back in the early 2000s. Before that I enjoyed Lubuntu with LXQt, but the latest iteration made some changes I didn't like, so back to the old familiar XFCE I went.

 

Both Xubuntu and Manjaro will run HQPlayer and, I assume, Roon. They also run the Qobuz streaming app via use of Nuvola Apps, based on Flatpak, another "container."  https://nuvola.tiliado.eu/app/qobuz/ubuntu/disco/

 

A central resource for summary overviews of, and links to, a very large number of Linux distros (as well as BSDs and other OSs) is Distrowatch: https://distrowatch.com


One never knows, do one? - Fats Waller

The fairest thing we can experience is the mysterious. It is the fundamental emotion which stands at the cradle of true art and true science. - Einstein

Computer, Audirvana -> wi-fi to router -> EtherREGEN -> microRendu -> USPCB -> ISO Regen (powered by LPS-1) -> USPCB -> Pro-Ject Pre Box S2 DAC -> Spectral DMC-12 & DMA-150 -> Vandersteen 3A Signature.

Share this post


Link to post
Share on other sites
36 minutes ago, Jud said:

First, thanks for the very thorough review. With the enormous number of options even a review of this size must leave out a lot of information, but there is plenty enough here to be very useful.

Thanks for the kind words, Jud.  As you observe, there's so much more out there that it seemed both impossible and unnecessary to present anything beyond the few systems and instances that seem most useful and suitable for audiophiles.  Remember that most people were both frightened of and frightened away from Linux as recently as 5 to 8 years ago, so any distro that's harder or more intense to use than Windows or MacOS will be a hard sell.  I seem to recall a post by Chris some years back in which he expressed doubt that audiophiles would ever adopt Linux 😉

 

Even today, there can be stumbling blocks with Linux.  The most common are network file shares (which most often means Samba issues), USB and other bus issues (which are often driver issues), and Wifi.  But Ubuntu, Mint, and a few others are now as easy to use as Win and Mac

 

36 minutes ago, Jud said:

Both Xubuntu and Manjaro will run HQPlayer and, I assume, Roon

Only Roon server and Roon Bridge are available for Linux.  Server is only available for 64 bit architecture, which is fine with me since all current major distros are also 64 bit - but there are still 32 bit Linux users with older machines that won't run Roon server.  You also can't run Roon Bridge on a Pi Zero, but it's great on 3s and 4s.

Share this post


Link to post
Share on other sites
1 hour ago, fas42 said:


I agree with this ... when the SQ is 97% or so there, every annoying deficiency of the playback system keeps on nagging one;

So with tongue only halfway in cheek, please help me to understand this. On what scale does one measure SQ? Is it a linear function? What’s the unit of measure? With what do you measure it? And, most importantly, if you’re already at 100% (which must be the case, unless you’re either suffering mightily while using your own system or not listening at all), will you be using your current system forever?

Share this post


Link to post
Share on other sites

Fantastic article. Just informing fellow readers that the best value proposition (in my opinion) is the use of x86 thin clients, whether NOS (new old stock) or used. Most allow for the use of Linux ( I use Lubuntu 12.04) and Foobar2000 with the UPnP component works extremely well as a renderer under WINE. Thin clients can be had for as little as $30 CAD. Those choosing to remain fully Linux can choose between gmediarender and MPD (or Mopidy)/Upmpdcli combination.

 

Happy addendum. Tried to use vst addons successfully with FB2K under WINE. Just add Foo_vst to the components folder on the WINE c drive and it should find any other vst addons placed in the components folder.

 

Rmel66.

Share this post


Link to post
Share on other sites

I use Audiolinux in ramroot for my two Roon endpoints. On my server I have tried Windows 10 Pro, Windows 2012, Windows 2016, Windows 2019, and Audiolinux ramroot. On all of the Windows and Windows server implementations I used Audiophile optimizer which I highly recommend.

 

So far the Windows Server versions have all sounded the best, with Windows Server 2019 to be out ahead of all others to my ears. I love Audiolinux on my endpoints but found it dry and bright when used on my server.YMMV.

 

SQ is what you decide it is for your system, in your space and for your own ears... unless you have a spouse/partner and then it gets tricky! All of the elements of your system combine to produce it.

 

.



"Don't Believe Everything You Think"

System

Share this post


Link to post
Share on other sites
1 hour ago, mourip said:

I use Audiolinux in ramroot for my two Roon endpoints. On my server I have tried Windows 10 Pro, Windows 2012, Windows 2016, Windows 2019, and Audiolinux ramroot. On all of the Windows and Windows server implementations I used Audiophile optimizer which I highly recommend.

 

So far the Windows Server versions have all sounded the best, with Windows Server 2019 to be out ahead of all others to my ears. I love Audiolinux on my endpoints but found it dry and bright when used on my server.YMMV.

 

SQ is what you decide it is for your system, in your space and for your own ears... unless you have a spouse/partner and then it gets tricky! All of the elements of your system combine to produce it.

 

.

The hardware also matters.  Everything from the cooling, mechanical rigidity and vibration isolation of the case & parts to the array and arrangement of components inside the box to the quality and integrity of connections among them can affect SQ. I suspect that these factors could also affect the relative performance of different software in the same box.  And this is as true for the server as it is for the endpoints - basics matter.  Some computer layouts and parts combos may even be more sensitive to software than others.

 

As for the spousal factor, I can only offer thanks for a wife of 47 years who’s tolerated speakers bigger then she is, component racks in the living room, and wires all over the place despite never having listened to music at home by herself on anything more sophisticated than a radio or TV 😁

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now

×
×
  • Create New...