Jump to content
  • bluesman
    bluesman

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

    PART 1: OPERATING SYSTEMS FOR x86 BOXES  [SBCs, ARM etc will follow in part 2]


    INTRODUCTION


    Hopefully, the last installment in our VPCA series got you thinking about what you want in and from the software you use to play your music files.  The next link in the chain is the device that turns a digital audio file into a dynamic bit stream (for subsequent conversion to audio frequency signal output by other devices that are further downstream).  We’re following the system roadmap I outlined in the first part of this series, and are now halfway through the front end:


    We start with retrieval of static archived files and their conversion to the clocked, dynamic data streams that serve as input into the system. We look at front end hardware and software in separate articles, installing, evaluating and comparing several of the open source software players you can install on current devices like SBCs.


    The last article looked at player software, so it’s now time to explore the operating platforms on which our player / renderer software will live and work.  There are many hardware options from which to choose, but the first choice is purely binary – x86 or not?  The x86 platform is well established, and you can choose from among thousands of sizes, shapes, configurations, and options.  You can run Win or Linux on almost any x86, and its footprint can be as be as a gaming tower or as small as an Intel Compute Stick. If you want serious computing power, you’re probably looking at big boxes and the only choice affecting your hardware selection is your preferred OS.  If you want MacOS, you have to use Mac hardware or suffer the consequences of a non-supported platform and a bootleg OS.


    You can almost always find a very nice, very capable x86 box for $300 or less from any of the major outlets.  You can buy computers without an OS and save some $ if you want to use Linux.  But Win10 Home comes installed on many of these loss leaders, and it’s fine for music.  I bought my wife a current production i3 HP laptop last week from Microcenter that has 4G of RAM, a 1T HD, gigabit ethernet, USB3, and 802.11ac wifi -  for $279.  It’s more than capable of making some seriously fine music and you shouldn’t hesitate to go this route if you want or need an inexpensive laptop as well as a music player.   These are not well suited to serving up and managing huge music libraries, so external storage and server (e.g.NAS) is close to mandatory.  But such systems meet the needs of many an audiophile.


    Your needs may favor a single board computer (SBC) in a tiny box.  SBCs are now a viable option for audio and almost everything else a computer can do – and these are GREAT value!  A Raspberry Pi 4 with 4G of RAM is an amazingly powerful little box that will run some great players from simple MPD to a full instance of JRMC - for $55 plus a power supply.  SQ is excellent, and it takes less than an hour from the time you walk in your front door with it to be playing music.  Those of you who built Chris’s Beaglebone Black player back in 2013 know first hand how fine even this years-old design can sound.


    Rather than address the big v. small decision now, I thought it’d be more productive to first explore what you can do with and get from available operating systems.  In general, you’re going to need an x86 platform for serious processing.  If you plan to use DSP, transcoding, upsampling, full screen high res graphics displays, etc to listen to DSD files, you’ll be caught short with less than a very serious machine.  SBCs can play DSD and MC very well, but they can’t handle a full complement of additional processes.  So, for example, Roon users would run Bridge on networked SBCs as endpoints, but the server or core would be on a more serious machine (mine’s on a NUC).


    I’m drumming up a decision support chart for the final installment of this series that will package everything we’re discussing into an algorithm or two.  Until then, our focus is on building a knowledge base from which to design and build a great system that meets a specified set of wants and needs.

     


    TERMINOLOGY AND GENERAL BACKGROUND INFO


    Keep in mind that the terms renderer, player, and endpoint have different definitions  and suggest different functions, depending on the frame of reference around the discussion.   Microsoft uses one set of definitions, DLNA uses a second, and various AS members have posted yet a few more.  None is right or wrong, but it helps to be certain that whatever term is used is understood  by the reader to have the meaning intended by the author.


    In DLNA terminology, front end devices are called renderers if they process files pushed from and controlled by a source (e.g. a cloud or onsite media server) and players if they pull those files from a source and control playback themselves.  Obviously, this distinction is a bit arbitrary, as the same device is often both a renderer and a player.  The control point is the user interface through which files are selected and directed from the server to the renderer / player in a format usable by the latter, and by which “mechanical” playback is controlled (e.g. play, stop, forward, next track, etc).  Although the classic control point is a mobile app, many renderers, players, and combos incorporate control points.  


    Control of playback gain / volume is controversial and not universally an integral function of the control point.  There is controversy over whether bit perfect playback of the input digital stream is audibly degraded by digital attenuation, and many still believe that analog volume control after digital to analog conversion should be used to optimize system performance.  There is no single answer to this question, as digital gain control has different effects on bit stream fidelity depending on design & quality of the hardware plus file format, bit depth, and resolution.  So I’ll leave this one alone, as it’s a popular accelerant for web forum flaming.


    Microsoft has a different lexicon for computer audio.  They define a streaming audio renderer as “...a media sink that renders... a single audio stream”.  They also define a device called an audio endpoint, which is “...a hardware device that lies at one end of a data path that originates or terminates at an application program...[and] either renders or captures audio. Examples include speakers, headphones, microphones, and CD players”.  As they provide no definition of a CD player in this documentation, we don’t know if they mean a transport or a more complete device that may include DAC, amplification, and an onboard speaker.   So confusion is understandable when using the terms player, renderer, and endpoint without defining them clearly.


    I strongly recommend that you go back to Chris’s excellent outline of the various hardware pieces in a computer based audio system in the UPnP world, as this is a god basic description on which we can build.  Although he wrote it in 2013, it’s still a valid presentation of general structure and terminology that helps clarify some of our current discussion.

     


    GOALS FOR THIS INSTALLMENT


    The main task we address here & in the next chapter is choosing, building, and using a platform for your computer audio front end that fulfills as many of your wants and needs as possible within your constraints.  Referring to the “archives to ears” flow diagram I provided at the beginning of this series, we’ll ignore data storage for the moment because it’s a huge topic on its own.  I’m working on a review and comparison of alternative devices and settings and will defer discussion of that stage to a separate article in the future.

     

     

    signal_path_frontend.jpg


      

    So for now, we’re bypassing data storage and focusing on converting a static audio file to a dynamic flow of data (for subsequent conversion to an audio signal further down the chain).

     


    THE PARADIGM (I.E. THE ASSUMPTIONS & EXPECTATIONS) DRIVING THIS WORK


    All software and devices have been used in OEM form exactly as you would get them

     

    • No alterations have been made to any components or code, with a few stated exceptions
    • The rare modifications made because of popularity and ubiquity are so noted and contrasted with the stock form and format of the entity under discussion

     

    Every setup was configured for best sound quality, bit perfect playback, etc

    There are so many tweaks, plugins, modifications, etc available for everything discussed herein that it was impractical if not impossible for me to address mods. This work provides a baseline on which you can build to your heart’s content, and I’m sure there are a million things you can do that would change your and my evaluation of anything I discuss.

     

    • Do not berate me for not trying something.  Instead, try it yourself and let us know how it works out.

     

    I have tried hard to be neutral and impartial, in order to bring you information you can use no matter what your preferences and needs.  My likes and dislikes are clearly stated wherever they might affect my judgment, and I tried hard to bury them for the duration of this evaluation.

     

    I personally acquired, built, installed, configured, and otherwise am responsible for the state of everything discussed here.  There are absolutely no real or potential conflicts of interest, and all opinions expressed are entirely my own unless presented as a quote.

     

    I am not a fanatic about any hardware or software and have no vested interest in any product, manufacturer, vendor, or other entity connected in any way with anything discussed in this work.

     

    Any difficulty I may have had with anything is most probably my own fault, as there seem to be very happy users of pretty much everything I’ve discussed.  But as I have more than a little knowledge, skill and experience, difficulties I encounter are probably going to afflict at least some of you too.

     


    COMPROMISES YOU NEED TO CONSIDER


    Your own wants, needs and circumstances will dictate how you approach this.  For example,


    What do you have now?

    • Are you happy with your current computer(s) or do you have to stick with it / them because of investment in / dependency on apps etc?  
    • Do you have an older computer you can repurpose for music?
      • Can it be refreshed to meet system requirements for your desired OS and music software?
      • Are you capable of disassembling and upgrading it as necessary?
      • Is it in good enough shape to justify investing in a new SSD, RAM  etc?

     

    Are you a tinkerer or willing to become one?

    • Do you have or can you get the necessary tools and equipment plus a clean work space?

     

     

    How developed are your skills at

    • computer setup, management, troubleshooting etc?
    • working without a GUI (e.g. command line)?

     

     

    Are you willing to adjust & configure during use or do you want your system on autopilot?

     

    What are your resources and constraints on them, e.g. space, budget, appearance, use by non-audiophile and/or computer-naive family members?

     

    What will be your system architecture and planned use?

    • single box between server and an analog audio system?
    • network player to feed multiple audio systems?
    • WAN streaming?
    • other?

     

    What do you want in the box – renderer? control point? Player?

     

    What kind of listening will you be doing with this device / system?

    • Intense, critical listening?
    • Background throughout the home?
    • Mixed use, e.g. some combination of mp3 pop, background, and serious DSD / MC among you and your family members?


    Based on these and other considerations that you’ll discover as you go, there are multiple approaches to consider, including

     

    • buying, building &/or assembling an entirely new box for dedicated music use
    • buying, building &/or assembling an entirely new computer for general / other use, including music
    • upgrading an existing computer for dedicated use as a music player
    • upgrading an existing computer for general / other use, and adding music
    • adding a second operating system to an existing computer as a dual boot for music
    • embedding a music system in a capable NAS to feed LAN / WLAN renderers & end points
    • adding a standalone music player with JEOS on a USB drive for use in an existing computer
    • leaving your computer(s) alone and buying a commercial audio device (streamer, network player, etc)


    What you want and what you need may not be entirely confluent.  I’m wedded to detail-rich interaction with my music.  So the interface between me and my music files contributes at least as much to my joy as does any other systemic factor, and I’d gladly change my music platform to accommodate software that thrills me well beyond what I use now.  Any advance in library presentation & management, information gathering & delivery, artwork display, sound quality etc that delights me and is within my budget would prompt me to meet its system requirements in a new dedicated box.  You may have other priorities and/or constraints that make that undesirable.


    Thankfully, I have the luxury of separating my music systems from the rest of my computing needs and functions.  Those who are limited for any reason to one computer for multiple needs may not be so flexible, in which case the choice is among 

     

    • players that run on the legacy system
    • adoption of new hardware and software for everything
    • a USB-based player with its own OS
    • a dual boot (if the existing platform is willing & able) 


    If you have a current computer with excess capacity and great specs, there’s no reason not to drop a music player into it, as long as SQ is acceptable to you.  Whether an optimized traditional Win, Mac, or Linux box running multiple simultaneous programs and processes can pump out top SQ worthy of an excellent audio playback system is a bit controversial.  There’s considerable debate over whether and how the compromises of a general use computer can affect it through factors like power purity, internal noise, quality of the interface through which it will drive the DAC, the battle among competing processes for CPU use, etc.  I have yet to see a data driven analysis of any kind that helps resolve this dichotomy, and I search often for something like a well controlled comparison or a set of comparative measurements with real world relevance to SQ.


    Of course, there’s also debate over whether a good general use computer as above is any more limited in SQ than an inexpensive device like a Raspberry Pi set up as a dedicated music machine.  We’ll explore this  further into the article.  Based on my own experience with over a dozen varieties of Pi, ‘bone, Chromebox, etc, I’m firmly of the belief that a well set up SBC with a good music app (player, renderer, endpoint, or whatever you want to use and to call it) driving a modern, well designed and executed DAC with files sourced from NAS or an external drive is quite capable of top tier SQ comparable to some 5 figure commercial devices.  But I also get pretty fine sound from my own Ryzen based HP running Win10, driving a few of my DACs. I suspect that driving multiple LAN & WLAN endpoints from a busy multiuse computer might affect SQ, but I haven’t investigated this because I’d have to move my best DACs, amps and speakers from room to room for any kind of a decent comparison.

     


    PRIORITIZING CHOICE AMONG OS, SOFTWARE, AND HARDWARE


    If I had to use one computer for everything, I would not migrate to new hardware and a different OS just to accommodate a music player, given the many fine players now available for Windows, Mac, Linux and its many variants.   I’ve chosen my OS first for many reasons on many occasions.  I used a Kaypro back in the early ‘80s because it ran CPM, which let me put together a suite of excellent (for the time) management tools for my medical practice.  When MIDI became practical for musicians (~1983), I assembled a Mac system around a 512 and set up both a MIDI keyboard with controller and a rudimentary synthesizer – and I used them in live performance as well as recording.  I added an SE and one of the first PowerBooks to the arsenal as they (and better MIDI interfaces) became available, while continuing to use the Kaypro in the office for a few more years.


    In 1998, I went to business school and had to get a good laptop. This meant Windows, since almost none of the software used in my MBA curriculum was available for Mac, and the few apps that were simply didn’t measure up to the Windows versions from which they were ported.  And as the IT infrastructure and capability of the university in which I taught and practiced was expanded, Mac was largely ignored despite the fact that many of the research departments (and Biomedical Engineering!) were almost entirely Mac.  So the clinical and business operations were entirely Windows, and Mac was used only by those who loved it enough to put up with being ostracized by our IT department.  Curiously, the back end servers for our intranet and internet were either mainframes or Linux boxes through the turn of the 21st century.  I also had to switch to an iPhone (starting with the 4) because our communications systems were barely Android-tolerant until the last few years., but they successfully integrated iOS devices even while keeping MacOS boxes in the second support tier.


    Once I switched to a PC, I never went back to Mac.  Windows software and peripherals became better and better for media creation and management, so I really had no need to switch. A company called Native Instruments brought out some amazing instrument emulators, the best of which (in my opinion) was a Hammond B3 app called B4.  I loaded it on my Windows laptop from business school and began to use it on the road instead of the ratty keyboards and organs in the backlines of most “lesser” blues and jazz clubs & venues. It sounded so good that the band asked me to use it instead of the real B3s and Leslie cabinets occasionally found at a gig.


    So I’m a great example of one who picks hardware and software to accomplish specific tasks.   I ran multiple rigs at the same time because no single platform would do everything I needed.  I never minded maintaining multiple systems despite occasional trouble remembering what did what on each one. I just picked the best OS, hardware, and software for the task at hand. This is a luxury is not available to everyone, for a variety of reasons – especially when the only thing your current system won’t do for you is play and manage a high quality music library well enough.  So this article looks at multiple operating systems both for standalone music platforms and for multipurpose use, in the hope that it will help those on the fence in deciding between a legacy box for everything and a low cost, simple second device as a dedicated music platform.  


    The choices today among OS, software, and hardware are not so clear cut as they were back in the days above because you can now do a lot of things equally well on multiple platforms.  I could only get good MIDI stuff for Mac when it first came out – Windows was nowhere. Native Instruments B4 was only available for Windows when I first got it.  And today, Linux and its many variants will do pretty much anything I and most of you need. So choice is freer and easier because you can now pick what’s most important to you and go from there on almost any common platform.  Further, the gross differences in look, feel, and function among the different systems are largely gone now, with all 3 coming much closer to each other than ever before.


    Given the cost of a new Mac and the cost plus inherent restrictions in moving to a new Windows box (e.g. version-specific networking issues, the lockdown of Windows 10 S, etc) , more and more former Windows users are adopting Linux, as are users of older Macs.  Some Linux distros are quite user friendly – my wife, whose picture adorns dictionary definitions of technophobe, has been using Ubuntu since her Vista box died some years ago and has actually gotten more comfortable with it than she ever was with Windows.  She started with Ubuntu 16.04 and is now on 19.10 – it may not be your grandfather’s Linux any more, but it’s now sufficiently good and friendly to be your grandmother’s Linux :)


    The bottom line on operating system choice is that you have to be comfortable with yours – you can’t be fearful or unhappy every time you log on.  It has to meet your needs and please you in the process.  It has to let you do what you want to do, and do it as you want to do it.  If you’re flexible, you can find OS love in a variety of packages – but none is perfect, so this is another of those compromises you have to make. The more set in your ways you are, the less flexibility you have in choosing an OS.  So, as with players, you really have to try a few to see what you like and what you don’t.  The sad but all too common alternative is to settle for the OS loaded on the bargain machine you buy for reasons other than informed choice (e.g. price, availability etc).

     


    THE SECRET LIFE OF OPERATING SYSTEMS


    Although the structure and function of an operating system are a deep dive for audiophiles, a high level working knowledge of the subject can be useful for us.  I suspect most of us have little or no idea what’s under the hood in structural terms.  Because the platform on which your audio system is based could affect usability and performance, I thought an introduction to the subject would be warranted.  I plan to do some extensive testing of systems beyond Windows, Mac and Linux for an upcoming piece once I master the new stuff myself and see how it performs in a few audio systems.


    Let’s start with some definitions and facts about our systems that you may not know:


    A kernel is a system program that controls all other programs on a computer

    • It’s the interface between the software and the computer’s hardware
    • It’s the core of an operating system & communicates directly with the hardware
    • It’s the first program to load into system memory & it remains there until shutdown
      • Other parts of the OS remain in storage until put into use
    • It manages memory, processes, tasks, and discs
      • It creates and removes memory allocations to facilitate software executions

     

    An operating system is a package of software and data that manages a computer’s resources

    • It’s the interface between the user and the computer’s hardware
    • Each of the operating systems we know and use has a kernel at its core
      • There are operating systems that do not include a kernel
      • Most embedded systems are in this group because they only have to provide a very few functions and can load & run everything they do as a package
      • This could be a rewarding approach to audio!
    • There are new and different platform structures that are no yet in common use by audiophiles but may offer advantages
      • Approaches like Unraid and Docker run on current kernels but create new kinds of structures that let you run packaged apps in “containers”
      • a container is a lightweight, standalone, executable package of software that includes everything needed to run an application: code, runtime, system tools, system libraries and settings
      • containers may be a solution to some of the problems of running all music functions in the same box with less demand on resources
      •  Unraid apparently began as a novel approach to NAS. Its current state could well lead to a better way to store and serve music files.
    •  A process is another name for a program that is currently running
      • Many processes are running in the background and are therefore inapparent except in the effects they may have on audio & other processes you do care about
      • Shutting down background processes can improve SQ and music app performance 
    • There are many kernels for Linux that are optimized for a specific purpose
      • Low latency kernels are just that – they reduce processing time to minimize delays in output relative to input. This is important in audio and video production, editing etc
      • Low latency kernels prioritize certain tasks (e.g. audio) and often slow down other processes because of this. So a low latency kernel may not be a good idea for an audiophile whose player is on a computer doing other things at the same time.
        • Modern multicore processors are much less affected, so a low latency kernel is more useful on a less sophisticated, less powerful computer than on a powerhouse
    • Windows kernels are much more standardized, with far fewer options than Linux kernels
      • Windows approaches functions like latency with path and driver modifications
      • ASIO is a proprietary Windows audio device driver protocol that bypasses any processing in Windows & directs the digital audio stream directly to audio devices to reduce latency and optimize SQ
      • WASAPI (Windows Audio Session API) is another & newer such Windows protocol
        • It consists of 9 interfaces that initialize and clock an audio data stream, write data to an endpoint buffer, configure & monitor control parameters, gives client access to session controls, and mediates several volume parameters
        • WASAPI can be bit perfect if configured for exclusive access to the sound device and there are no digital gain controls in the processing path
          • Windows volume controls are bypassed in this mode & output to DAC is at 100%
    • Here’s a list of Windows kernels with tech details for those who want to know more


    There are different types of kernels, e.g. monolithic and microkernel.

    • Microkernels are smaller, slower, and more easily extendible than monolithic kernels. They contain more code and (unlike monolithic kernels) do not take down the system if a service crashes.
    • Unlike monolithic kernels, microkernels store user services and kernel services in separate address spaces.
    • Linux uses monolithic kernels; MacOS and Windows kernels have features of both;


    Here are some common OS myths and the truth behind them:


    MYTH: MacOS is Linux or Linux based. 

    • When Steve Jobs left Apple in 1985, he founded a new company called NeXT and rushed to get a new OS and computer to market. So he and his team used the Mach kernel developed at Carnegie Mellon and a good deal of BSD code developed at UCal Berkeley on a platform built to some degree from Unix pipes – he called this the NeXTSTEP platform. 
    • He couldn’t make that viable, sold NeXT to Apple and returned with it to make NeXTSTEP the base for MacOS.
    • The MacOS kernel is known as the XNU, which stands for “XNU is Not Unix”.
    • Linux kernels are newer and written for it; the XNU kernel is based on some old stuff.

     

    MYTH: The Windows registry is the main reason it’s so prone to crashes, hacks, etc; MacOS and Linux are better because they don’t have a registry.

    • Windows, Mac and Linux all have to store operating parameters between reboots so they can initialize drivers and bring the system to life.
      • Windows uses a central registry that is accessed by its software as needed to gather configuration data
      • Mac and Linux have configuration utilities like sysctl that use named text files just as Windows uses text strings called registry keys to control kernel parameters
      • The registry is a kind of database; so are the sysctl files; current design trends are moving toward XML instead;
      • It is just as easy to mess up a Linux or Mac config file as it is to do so in the Windows registry; 
    • The Windows registry does become full of bloatware data that can be very hard to find, let alone eradicate. But in and of itself, it’s not as terrible as many make it out to be.

     

    MYTH: There is an operating system that offers  demonstrably better SQ than any other

    • The differences in SQ among top quality music systems using Windows, Mac and Linux are simply not sufficiently large or consistent to make any one a clear winner

     


    PUTTING A  PACKAGE TOGETHER


    This work will explore the wide world of computer architecture for audio use in 2 parts, including operating systems  and hardware plus the rules for their interaction (which determine what a given system can do).  We have to look at hardware (in the next installment) specifically because of the growing trend toward OS specificity for processors and their environments.  If you go the virtual machine route to run a music system from a legacy computer with some miles on the odometer, you may encounter difficulty with either Intel or AMD chips (or both, depending on your choice of virtualization platform).  These incompatibilities are gone in the latest versions of both makers’ chips – but you have to activate virtualization in some systems.  


    Peripherals like networking adapters can cause serious headaches, as even a simple WiFi dongle may give you fits trying to get it to connect on Linux.  Configuring an OS for optimum SQ also takes a bit of knowledge and effort on some machines.  My last rack mount recording rig included a wonderful 18 channel TASCAM digital audio interface – it was 24/192 when that was high res, and it sounded great both for recording and playing back.  My DAW was on a Win7 PC that had seen better days, so I finally “upgraded” it to the then current LTS (long term support) Linux distro……...and nothing I did brought the TASCAM online.  After weeks of unsuccessful searching for drivers, support, and divine inspiration, I sold it to a friend and bought another interface specifically stated by the manufacturer as being compatible with Linux.  


    There are over 100 different Linux variations , adaptations, derivatives, and clones available today.  Some are easy to download, install, configure and use while others require more knowledge, skill and experience than the average audiophile brings to the game.  Some will run well and reliably on ARM-based SBCs like a Raspberry Pi, while others have more demanding system requirements. So integration is the key – planning will avoid 95+% of the problems that plague the unprepared.


    Of the 31 free or low cost players I evaluated last time around,  7 only run on Linux, 4 only on Mac, 5 (not including WMP, which I did not include in the review) only on Windows, 2 only on ARM based SBCs, and 1 only on x86_64 based mobos.  A few are packaged with their own JEOS (“just enough operating system”), most of which turn the hardware into a music-only device with little or no access to the underlying OS.  A few will boot, load and run from a USB stick and leave the “host” computer entirely unaltered and intact when removed.


    The intrepid Linux users among us can run most Win-only players on Wine with surprisingly little compromise, and you can always virtualize if you’re wedded to a single platform. If you need or choose to do this, I strongly suggest buying one of the full packages from VMWare, VirtualBox etc because audio players on the freeware packages I tried (VMWare, VB, and Boxes) were slow and cumbersome to the point of being essentially unusable.


    Players that run on multiple platforms are not necessarily identical across the board, and the differences can be so startling that you might love a player on one OS but hate it on another.  Look at the main comparison chart I included in the Front Ends article and you’ll see a few that I found to be great on one OS but unacceptable on another.  For me, DeaDBeeF was great on Linux and terrible on Windows.  Jajuk was easy and fun on Windows but tedious on Linux.  So if your priorities start with your player, you’ll need to decide on the best OS only after you’ve made your choice of player.  And you have to try each player on every platform on which it runs, to be sure you know what you’re going to have when you’re set up.  Had I decided to use DeaDBeeF on my Win10 machine because I loved it on a Linux box, I’d have been grossly disappointed.

     


    LIFE AFTER LINUX: GOING BEYOND THE BIG THREE


    Win, Mac and Linux(including SBC variants) are considered to be the top tier in operating systems for most non-scientific and non-commercial computing.  There are many other operating systems out there, some of which may offer advantages for dedicated use in sound production and reproduction – they simply haven’t been developed for audio yet.  Although I’ve downloaded, installed and played with several of them, I don’t use any of them for anything and am neither recommending nor reviewing any of these.  I just thought it important that we all recognize how much more is available to us than most realize.  And these are all great opportunities for learning and playing at no cost. As they each require you to configure your audio settings, you may find a surprising tweak or twist that you didn’t know about, as has happened to me many times.


    Let’s start with Illumos, an OS based on OpenSolaris (an Oracle product abandoned by them about a decade ago).  There are several flavors of Illumos, of which the one probably best suited to our needs is OpenIndiana in its latest release (called Hipster) with Plex or Kodi installed.  Then there’s Haiku, a 20 year old open source OS based on the long-forgotten BeOS written for the BeBox – and you get 100 bonus points if you remember this baby!  


    Ever heard of ReactOS, a de novo Windows-like system built on WINE code?  How about Tizen, Tails, or Qubes?  There’s  a very lightweight OS called Kolibri that occupies 3 megs complete with GUI.  It can read Fat, NTFS, and EXT files and there are many applications written for it.    I even discovered a FLAC player created for Kolibri by an enthusiast, and there are many posts on their forum addressing music use of this fascinating system.  Slax is a tiny, lightweight, live linux-based OS that resides in a USB stick along with persistent data storage, requires no installation, and can probably run lightweight players.


    If you have an old Apple / Mac /PowerMac or an Amiga, you can play with MorphOS (“a lightweight, highly efficient and flexible media-centric operating system”).  And if you like this approach but are running an x86 system, there’s always AROS, which is described by its team as “...a lightweight, efficient, and flexible desktop operating system, designed to help you make the most of your computer. It's an independent, portable and free project, aiming at being compatible with AmigaOS at the API level (like Wine)”.  Another oddball of some interest is MenuetOS, “a pre-emptive, real-time and multiprocessor Operating System in development for the PC written entirely in 32/64 bit assembly language” with an audio driver.  

     


    CHROME MAY BE DEAD ON CARS, BUT IT’S ALIVE AND WELL IN COMPUTERS


    No discussion of value in modern computing platforms is complete without including the world of Chrome.  Chromebooks and boxes are simple computers designed and built around Google’s Chrome operating system, which is now a decade old and uses a Linux kernel.  Although they only have about 4% of the home computer market to date, sales have grown at a rapid rate because they offer great value in hardware, software, and usability.  Tablet use in schools has dropped a bit while Chrome-based devices (‘books and ‘boxes) are increasing in popularity for students.  And the sophistication of Chromebooks has really soared with use of far more powerful processors than the tiny original heart of what began as “netbooks”.


    Functionally, the Chrome OS conjures up a modern interpretation of the old dumb terminal – mainframe relationship.  Cloud computing replaces the mainframe in this equation, but the work and data flows are very very similar in that the C-book or box is essentially the GUI for a “back office” computing facility.  They boot fast, work cleanly and spend most of their time and energy carrying your input to the cloud and returning your requested output to you for display / playback or further input.  I suspect that some serious development could produce excellent music players that would source or stream files both from the cloud and from network storage set up as a robust media server.  I haven’t yet found a Chrome extension (which is what they call apps because the entire GUI is browser-based) that would play lossless files, and SQ of the Chrome players I tried on my Asus Chromebox is not great.  But there’s some potential for the value conscious audiophile who’s willing to invest a few hundred dollars and some time learning how to make this work.


    FWIW, I did a double boot install of Ubuntu on an original Asus Chromebox using Crouton.  This worked OK and made some fine music, so I went the extra mile to eliminate as much extraneous activity as I could.  This meant purging the entire Chrome package, after which I tried a few different operating platforms with fine results.  Right now, it’s running Ubuntu 18.04 LTS with Roon Bridge plus my development ColdFusion web server and a MariaDB backend.  It can always be wiped clean or restored to its OEM Chromebox status running the ChromeOS.


    Chrome is for Chromebooks (and ‘boxes) – but it ties you to Google from log-in to services.  You could put NayuOS on your Cbox to preserve the Chrome experience but leave Google behind.  And if you like the way Chrome operating system works and feels but want to have it on a “real” computer, there are a few Chrome-like systems that will let you do that, e.g. Remix and Phoenix.  There’s also a Linux distro (based on Xubuntu) called GalliumOS that’s specifically for Chrome devices (e.g. it has an embedded Chrome mouse driver).

     


    ANDROID


    Google’s been a busy boy!  Their Android OS started as a mobile platform, with Chrome OS coming out later as their “desktop” system and intended only to be embedded in Chromebooks and ‘boxes.  But Android and Chrome have a lot in common, and their paths are crossing in more and more places as time goes on.  With equal parts effort and luck, Android can be installed on a Raspberry Pi, although Google’s not supportive and I went through 4 different schemes before finally getting 8.1 (“Oreo”) to work using this YouTube tutorial.  It has to be done with unofficial 3rd party versions, it’s not exactly simple or foolproof.  Worse, the version I got to run did not include the Google Play Store – so you have to figure out how to add it via APK, which is also not easy in devices that were not originally Android.  And without the Play Store, you have to find and install your music players as APKs.


    Many Android apps will now run on ChromeOS, and I’ll eventually get to trying the best music players on one of my Chrome devices.  There are a few really nice players out there now that have proven themselves on Android phones, and that potential is the reason I’ve included Android in the discussion.  High quality players like Poweramp, Onkyo, Neutron, USBAudio Player, and Foobar (the very same!) do a very nice job on current Android devices and should be just as good on capable Chromeboxes and similar devices running Android (once we figure out how to install and set them up easily).  I can see how a Chromebox or ‘book with a decent CPU and adequate RAM could be a pretty fine music player for very little money. 

     


    MOBILE PHONES FOR AUDIOPHILES


    Don’t discount phones for audiophile use.   A few, like the LG V10 and HTC One A9, have acceptable onboard DACs for general listening.  More importantly, many phones and other devices running Android 7+ will output digital audio to a USB DAC - but you have to check each one to be sure it’ll work and to find out exactly how to set it up.  Many, but definitely not all, will require (but do a good job with) an OTG dongle driving a USB DAC.  Again, you have to check each model specifically to be sure it’ll both put out a digital audio bitstsream to a USB DAC and to find out exactly what connectors are required.


    With a top quality player on it, a good mobile phone can serve you well.  I didn’t look at players for phones for the front end article in this series because it wasn’t relevant, and I’m working on a mobile sound piece now.  But if you want to use your phone for serious audio, look at players like Neutron (available for iOS, Android, and Windows), a fine and inexpensive app that will let you configure bit perfect playback on some platforms, play native DSD / DoP, Chromecast directly to the device, and much much more.  Here’s a nice overview of USB audio with special reference to Android.

     


    THE MAJOR CONTENDERS & THE FINDINGS


    Nota Bene! If you skipped or forgot the content of the “compromises you need to consider”section above, I’d suggest you review it before making any decisions about your proposed system build or change.  You probably won’t please yourself if you don’t put your own needs, wants, and idiosyncrasies first – and you’ll have to prioritize those three first for the highest probability of success.


    These are just some of the factors you’ll have to value and prioritize for yourself before you can design and build / install a system that will fulfill your wants and needs well:

     

     

    user interfaces

    • hardware
      • functional (e.g. physical controls)
      • aesthetic
    • operating system
    • players and other music apps you’ll use regularly
    • audio configuration settings and tools
    • music library displays
      • associated art and information
    • other apps you want, even on a dedicated music device

     

    system interfaces

    • audio
      • USB, S/PDIF, AES etc
      • optical vs coax
      • balanced vs unbalanced
    • network
    • wireless vs physical

     

    functionality

    • single vs multiple endpoints
    • desired fiddle factor (set & forget? gain only? EQ? DSP? Playlists? Etc)
    • scalability (e.g. MC, higher res, new formats)

     

    number and nature of boxes in the system

    • dedicated music vs multipurpose device
    • single box vs “network streamer” into sound system vs discrete physical stages
    • analog / digital structure of the system
      • fully digital from player to output stage (“power DAC”)
      • analog from DAC output

     

    desired and/or tolerable level of complexity (which most often also equals flexibility)

    • total cost (including licenses, updates, all prerequisites for your chosen OS, hardware etc)
    • how much of the initial outlay is recoverable through resale in case of remorse 
    • number, sophistication, tolerance and willingness to learn of intended users
    • ease of troubleshooting and solving problems
    • support for each and every element of the system
    • likely need
    • source, availability, and response time
      • one of my DACs has to be sent to China for warranty service!
    • cost
      • nature and extent of any warranty coverage
      • effort required to get it

     

    your own levels of flexibility and willingness to compromise

     

    severity of your personal case of gear acquisition syndrome

    • how likely and how often will you update, change, or otherwise mess with your system? 


    It’s amazing how helpful it can be to list all of the factors relevant to you in two columns, prioritized for want and for need, in a simple spreadsheet.  Once you have the factors prioritized, copy the list onto a separate sheet and replace the factor name with the actual item(s) that seem to best fit what you’ve chosen for the characteristic in question (e.g. your USB DAC of choice, based on price, features etc if you only want USB and know about how much you want to spend on it).  Add a column for cost sensitivity and one for the strength of “want over need”, and you’ll be looking at a great decision support grid.

     


    ANOTHER CONSIDERATION WHEN SELECTING AN OPERATING SYSTEM


    Remember that Windows and MacOS each have only one current level of development at any time.  Windows offers the current system in multiple releases that differ in features and cost, but Win10 has a Win10 heart whether it’s Home, Enterprise, Pro or other.  If you buy a computer with an older version of Windows on it, you can upgrade if the hardware’s up to it – but it’ll cost you.  The system requirements for Win 7, 8 and 10 are essentially the same, so you’ll almost certainly be able to run Win10 on anything that came to you with 7 or 8.  But the machine may not be able to take advantage of every feature and function because standards change, e.g. max RAM and bus speed.


    Apple has only one general release of its OS at a time, and it’s what you’ll get if you buy any new Apple computer – there are no options or choices as far as I know.  If you buy an older model, you’ll get whatever version of MacOS is on it when you buy it, but you can upgrade MacOS to the system requirement limits imposed by your hardware at no additional cost.  However, you cannot buy MacOS from Apple (the only legal source of current releases) unless you’re the registered owner of a Mac computer of some kind.


    Linux, on the other hand, is available in over 100 flavors, many of which are minor variants of each other but some of which differ greatly.  Similarly, there are many operating systems available for SBCs.  Some of them are distinctive and others are interchangeable clones of each other.


    If you’re buying or building a computer as a music source, your choice of operating system depends on the hardware you choose.  So you need to prioritize among hardware, operating system, and applications before you can design your system to know what’s most important to you.  If you want Mac hardware, you’re stuck with MacOS - and if you want MacOS, you’re stuck with their hardware.  If you want Windows, you have a variety of hardware platforms from which to choose.  But if you’re buying new hardware with an installed OS, you’re going to get whatever version of Windows is installed.  Look very closely at those bargains on the internet and at the major retailers, because they don’t all include the latest versions of hardware or software.  You may be getting that great price because it’s the last model with 10/100 ethernet, 802.11n WLAN, or an older OS.


    If you want Linux, you’ll be interested in the detailed Linux summary further on in this article.  Some versions are easy for novices to use, while others are both labor intensive and less than intuitive to configure.  Some have excellent GUIs for every function and are as easy to use as Windows and Mac.  But others require you to download and install features you’ve come to expect from traditional Win and Mac computers (or, worse yet, require use of the command line in a terminal window).


    If you decide to use an SBC for your music player, you’ll need the guide in that section a few pages further on.  Many of these little boxes have a recommended or standard OS that’s a great place to start, e.g. Raspbian for the Raspberry Pi.  Once you get set up and running, you can play with other operating systems and players as you wish just by downloading them and burning them on a microSD card.  Then you can power down and switch cards to change your operating system, player, etc and easily compare multiple alternatives one by one in the same playback system.


    And if your priorities start with the player, you can choose among all the alternative hardware and operating systems that will accommodate it.  For example, a full installation of JRMC will run on Windows, MacOS, Linux, and Raspberry Pi.  Earlier versions of the Pi were challenged by the JRMC GUI, although the SQ is quite fine.  If you put it on a Pi model 4, you’ll be amazed at how well it runs both visually and audibly.  If you run Roon, you can’t yet put the core on a Pi - you’ll need to run it on Win, Mac or Linux box.  But you can use SBCs as endpoints / renderers for stellar SQ.

     


    BITS MAY BE BITS IN A DIGITAL MUSIC FILE – BUT NOT IN AN OPERATING SYSTEM


    You’re undoubtedly familiar with the 32 vs 64 bit processor dilemma.  64 bit processing is not new –  IBM used 64 bit word architecture in the early ‘60s, and Crays were 64 bit computers from ‘76. IBM developed the x86_64 platform for consumer devices in the late 1990s, with both Intel and AMD producing chips for it within a few years.  ARM offered a 64 bit chip in 2011, and many SBCs are now 64 bit including the Raspberry Pi 4, which is light years ahead of the 3b+ (especially with 4G of onboard RAM). The 3b+ had 64 bit capability but was a bit weak with 64 bit software.  


    Most consumer computers sold after 2003 had 64 bit capability, even though not much 64 bit software was available yet.  The first OS in consumer machines to be fully 64 bit was MacOS Snow Leopard in 2009, and FWIW the first phone with a 64 bit processor was the iPhone 5S.   Even today, you’re offered a choice of 32 or 64 bit when downloading many apps.   But fewer and fewer software packages of any kind are still being offered for 32 bit machines, and 32 bit processors are dinosaurs that are following the same path to extinction in PC and Mac devices.  Many operating systems and apps are only available for 64 bit devices, and it’s only a matter of time before the end of 32 bit versions of new programs. 

     

    A 64 bit processor can handle much more data at a time than a 32 bit processor, and it can access 4 billion times the physical memory usable by a 32 bit CPU at the same speed and other operating parameters.  A 32 bit processor can only access 4G of RAM, which is in itself a limiting factor on its utility for major media work.  There are some OS limitations on memory use.  Here’s a list of memory size limits on Windows releases back to XP.

     


    WHO’S YOUR DADDY?


    No matter which platform you use and how simple your system, you’re going to be downloading and installing new programs at some point.  You’ll also have to update the ones you’re using, and you’ll have to do an occasional bit of troubleshooting.  These functions are pretty straightforward for Windows and Mac, because they control many of the sources of your programs.  When you use the “factory store”, you know you’re not an orphan.


    Microsoft has compliance standards that must be met before they’ll put an app in the Microsoft Store.  They even have a “Windows App Certification Kit” for developers to use to evaluate their own apps before submitting them to MS.  Apple has a comprehensive developer program and good support for those who want to sell their apps thrugh the iTunes store, including a beta program to let you offer your app free to up to 10,000 users and get feedback to improve your product before submitting it for consideration.


    You can get 3rd party apps on the internet from countless sources - but those without MS or Apple certification can be problems for you and your computer, and there’s a lot of malware lurking on independent download sites.  So you can keep it in the family with MS and Apple certified software.  You can even get some support from MS and Apple for software they provided to you, although it’s not as exacting and timely as we’d like.


    All freeware Linux apps are 3rd party because there’s no factory and no factory store. You need a paternity test to identify the lineage of a lot of the Linux software you download (and use at your own risk).  Some of it is so obscure that neither Ancestry.com nor 23 and Me will be able to help.  Many open source developers identify themselves and stand behind their creations to the point of providing quick and helpful feedback through online panels, forums, etc.  But it’s all voluntary unless you buy support.


    Linux apps are provided through data repositories called Personal Package Archives (PPAs) that are accessed either from the command line or through software management programs.  The well known and widely used distros have their own PPAs, with some oversight provided.  For example, Ubuntu has a process for getting your app into Ubuntu that includes volunteer reviewers, guidelines, etc if you’re willing to let them see your creation in development.


    The first and best place to start a search for a Linux app you want is the repository of the distro you’re running.  If it’s available through the included software manager, installation will be easy and updates can be configured just as they are in WinMac.  There are many fine music players and media management apps in the major PPAs, so you probably won’t have to stray to find one you like.


    But many Linux apps are provided from private PPAs, which lack consistency in oversight etc.  Obtaining software from an obscure and unknown PPA is a bit of a gamble, despite the paucity of Linux malware.  One bad package can brick your computer and send you back to square 1.  For audiophiles, the major private PPAs are no concern at all.  I gladly install JRMC and Roon from their private PPAs because these programs are not currently available through any repository associated with a Linux distro.  But using 3rd party PPAs usually means using the command line in a terminal window.  There are more and more installation packages now with scripts that execute the installation for you – but, again, loading an unknown script from an unknown source into your computer is a security risk


    The best way to avoid problems in this regard is to stick with the best known and most reputable distros.  You’re no more likely to get bad software from the Ubuntu or Debian repositories than you are from Apple or Microsoft.  Just be careful when you start adding accessories from unproven sources.  Apps from MS, Apple, and the major Linux repositories have traceable lineage.  You let a stray mutt into your computer at your own peril.

     


    LEGACY CONSIDERATIONS


    You can still make some fine music with older computers running 32 bit systems.  I run Ubuntu Studio on a 2005/6 Toshiba Satellite U205, along with JRMC and Roon Bridge, with pretty good results.  I tested many of the players in the last installment of this column on it, to see how they did with hardware and OS limitations – and SQ was quite fine with ALSA.  In the process, I also loaded both 32 and 64 bit versions of several operating systems to compare, and I heard absolutely no difference in 16/44 FLACs or in DSD64 files between 32 bit and 64 bit versions of multiple Linux distros.  The only “advantage” to 64 bit systems is access to a lot more RAM – so players that play from RAM could take advantage of lots more RAM instead of the 4 gig limit on 32 bit systems.


    Along with new stuff, we’ll address legacy hardware in the next installment, which is a serious shootout of computing devices new and old, fancy and simple, big and small, old (some over 20 years) and new.

     


    WINDOWS


    Windows in general


    The operating system everybody loves to hate is now 34 years old.  The current desktop family (Win10) is the 26th PC generation, and the current Windows Server 2019 is the 13th server generation.  The current PC version (1903, code name 19H1) is the 8th Win10 PC version since the debut general release of Win10 in July 2015.  It will be updated to version 1909 (code name 19H2) in a program that’s supposed to start this month – but more about that later.  And there’s a secret successor version codenamed 20H1 that was released in February 2019 to members of the Windows Insider program (in which those with a Win10 or Server 2016 license or newer) can get developer releases to try.


    For this review and comparison, I’m running  the latest build (1903) of Win10 Home on a 2018 HP Pavilion 590-p0085 (AMD Ryzen) that’s stock except for a second 2 HDD used only for selected backup.  For historical interest, I still have an AMD Athlon X2 64 bit Gateway laptop running Win7, and I started with an early version of Win10 back in 2016 on a slightly less capable HP Pavilion.  Although direct A-B-X comparisons spanning a decade+ are just not practical, I did compare the Win7 Gateway laptop with Win10 on my HP tower driving the same Emotiva USB DAC via USB2 with Foobar2000 as the source player. And in an unsurprising spoiler, I’ll just say right now that I heard absolutely no differences at all in SQ either between Win7 and any version of Win10 on the earlier HP or between Win7 and my current Win10 system, using Redbook FLACs.


    Win10


    There are at least 11 current versions of Windows 10, of which Home, Pro, and Education are probably the only ones of interest to us.  While you could run a music system from your business network on Enterprise, I suspect this is not an audiophile’s dream.  The main limitation of 10 Home for our purposes is that there is neither domain nor group management in the Home version. These may not be important to many, but I do use domains and group management for access to various network nodes like NAS, routers, switches, etc.


    We should mention Win10S, which is either an operating mode of 10 (per Microsoft itself) or a separate version of 10 (e.g. per The Guardian).  Microsoft says that it was designed for low cost computers, those used for education, and some premium devices like their new Surface laptop. They also say that the S stands for simplicity, security and speed – it’s spec’ed to boot 15 seconds faster than the same machine running 10 Pro and to run as fast on day 1000 of use as on day 1.  


    S mode adds MS BitLocker encryption system that’s not available on some other versions of 10, which might be desirable to some.  But it also limits you to installing apps from the MS Store and to using Edge for web browsing. So you can’t install and run many major apps unless you buy them through the MS App Store. I was very surprised to see what’s missing there, e.g. the only Quicken program I could find is Quicken Will Maker 2019. I suspect you can use the latest cloud versions of major software because they run through your browser.  But browser based music is not ideal under the best of circumstances – and I’m not a big fan of Edge. So you’ll have to download your players from MS.  Apart from Foobar, there are few good players on the list.  I saw none that come close to Foobar.  So S mode is almost certainly something we would all turn off immediately.


    You can turn S mode off in “settings - update & security – activation – Go To The Store”.  But it takes a little care, as you may also see a link of the same name under “Upgrade your edition of Windows”.  DO NOT click this if you just want to inactivate S mode!  Once you turn S off, you cannot reactivate it.


    In general, Windows is Windows regardless of the version you run.  They keep changing the system’s graphics and screen real estate on us, and you have to relearn where to find some functions when they update it.  I’ve had to Google a few things to find out where a toggle was hidden, and there are often alternative layouts available if you know how to change them.  


    As a university professor, I’m eligible to get the Education version of Win10 and plan to do so out of curiosity.  But I don’t see how it could impact this survey so I went ahead without it.

     


    WINDOWS FOR AUDIOPHILES


    This one is easy: I detect absolutely no difference in SQ among all flavors of Win10 set up and tested the same way, e.g. ASIO or WASAPI, minimal use of background apps and other loads on the CPU.  In fact, I don’t hear a consistent difference between Win10 and Win7 driving the same DAC and system with the same 16/44.1 source material using the same version of Foobar.  No, they were not on the same computer – my only remaining 7 machine is a Gateway Media Center minitower and my regular 10 machine is a 2018 HP Pavilion.  The power supplies are almost identical because the OEM unit in my Gateway died a few years ago and the easiest thing to do was plug in a new external unit – and I’d recently bought a few very good PSs on clearance “just because”.  


    Any improvements one might reasonably hear or otherwise experience are more likely to come from better hardware than from a 7 to 10 switch, assuming you buy a PC with Win10 on it to replace one that came with 7.  Although the basic system requirements are the same for 7, 8 and 10, upgrading from 7 or 8 to 10 on the same old box may not produce the results you want because of hardware limitations.  My boot time with 10 from an SSD is faster than it was with the original HDD (~5 seconds vs 8), but it’s probably more because of the SSD than the OS.  A published comparison using BootRacer shows boot times for 7, 8.1 and 10 on the same machine to be 5, 4, and 6 seconds respectively.  


    As 10 manages memory more efficiently than did 7 or 8, it should perform better at complex tasks requiring serious RAM.  But this is unlikely to affect most audio functions from SQ to management of all but truly huge libraries.  And more RAM actually lengthened the boot process in this test, with a 5.1 second boot time using a single 4G DIMM compared to 5.3 seconds for any configuration of 8G or 16G and 8.17 with 24G on board.  If you just plop a music player in a serious Windows box you use for intense computing for business, gaming, web hosting, etc, audio functionality like SQ, library and art loading etc could suffer if everything else is competing for processor, RAM, etc at the same time.  I did not test this and cannot find any published reviews – so it remains hypothetical.  But if you put big money into a Windows computer used simultaneously for a specific purpose other than audio, it makes the most sense to me to build or buy a dedicated audio device like a NUC or a Pi.  And if you truly love Windows that much, you can almost always find a capable little PC with 10 Home on sale for under $300 at any of the major computer outlets.


    After doing my own comparisons, I searched the web for any material that might reasonably refine or otherwise improve the utility of these impressions. and observations on Windows.  I found well written and supported summaries like this one


    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.


    ... I have been using Windows 10 with my other computers over the last few weeks and have not noticed any sonic difference to speak of among the various machines and different DACs (like the ASUS Essence One listening with headphones).


    ...and I agree completely.  If you are using Windows for audio because you want or need it for other reasons, fulfill those needs and the music will follow – Windows is a fine operating platform for audio and I can find no reason not to use it if you like it, with the above cautions and considerations.


    Neither I nor anyone else seems able to make an audiophile case for any version of Win10 over any other. From my own experience, I’d upgrade to 10 from earlier versions if my hardware was up to it.  I’ve had fewer glitches with 10 than any prior version – I haven’t seen a BSOD in years, and freezes or other operational mishaps have been rare to nonexistent.   


    The last question for Windows 10 is which version to get.  The main differences between Home and Pro are security related in the forms of BitLocker device encryption and Windows Information Protection.  BitLocker lets you lock down what’s on your device if it’s lost or stolen, and WIP is a kind of virtual firewall that isolates selected information on the device from the rest of the device and from networks to which it may be connected.  Neither of these is of any importance to me, but I don’t think they’ll impact audio performance assuming you set up access and permissions to network file shares properly.  Business functions like active directory integration, mobile device management, and group policy implementation are included on Pro but not Home.  And although I can’t actually tell whether 10S is a separate version or a mode in other versions, I’d inactivate it because it will limit your choice of players and other software…..and because it’s just wrong :)

     


    MacOS


    Wow – how do I address the big Mac?  Macs have always been presented as cool, contemporary, and very user friendly.  They’ve been touted since the beginning as superior to Windows PCs for artists, photographers, musicians, authors, publishers and others in the arts and media.  Truth be told, MacOS is almost as old as Windows – and even the latest version is built on code that’s older than Windows.  Steve Jobs’ return to Apple in 1998, along with their purchase of NeXT Computing (the company he founded in an abortive attempt to compete with them after his forced departure a few years earlier), led rapidly to the introduction of what we now call MacOS 10.X in 2000.  The back story of its development is summarized above in the secret life of operating systems.  


    Apple was in some serious trouble in the late ‘90s.  MacOS was born out of desperation and raised in a witness protection program so no one would know what it really was.  It was hardly a cutting edge creation when introduced, and it continues to bore software engineers and solution developers to tears to this very day.  Fortunately, the people behind the hardware, GUIs and functionality have been very creative and managed to turn some very tart fruit into a drink many people like.


    Many people were shocked and appalled that Apple actually sold a MacOS “public beta” for $29.95 in 2000 so they could get feedback from early adopters – talk about chutzpah!!  And since that initial release of what they called Kodiak, there have been 15 versions of the system to date, the most recent of which is Catalina (10.15).  The look and feel of Mac remains similar today to its presentation 20 years ago, and MacOS is a lot more Windows-like than prior Mac operating systems (which touted controversial items like one button mice without scroll wheels as being better because they were simpler).  Many of the traditions of Mac are now gone.  Finally, the desktop dashboard from 2005 is gone from the 2019 MacOS release (Catalina).  It was a relic from the good old days when Apple thought that skeuomorphism (i.e. making software look like real world objects on the screen) was cool. This blog post on MacOS is both entertaining and informative.

     


    Apple vs the Audiophile: hardware issues


    Fast forward to 2019.  In the world of Mac, it’s more like a slow forward – change has crept in on little feet, and (except for hardware design and construction) innovation is scarce from the user’s perspective.   MacOS still looks and feels pretty much like it did as a child.  Worse, Apple still locks down virtually everything so that it’s very hard to go Mac without a Mac - you still can’t run a Mac OS on anything but Mac hardware without some serious compromises that include breaking the law.  In fact, you can’t even get a real copy of the OS from Apple unless you’re a registered Mac owner – you’re limited to unauthorized (and illegal) versions.  


    The best known source of information for those who would try to run MacOS on non-Mac hardware is Hackintosh, whose website clearly states that “... Apple does not support Mac OS X on non-Apple hardware and installing Mac OS X on non-Apple hardware may be illegal in your country”. ‘nuf said.  If you want to run MacOS, you’ll save yourself a lot of aggravation if you simply buy Mac hardware.  I’ve run virtualized MacOS machines on a powerful Linux box using both VMWare and VirtualBox consumer freeware, and performance is unsatisfactory for audiophile use.  The paid commercial virtualization programs may be better, but if I’m going to spend serious money on this, I’d rather buy a Mac.  It makes no sense to me to virtualize MacOS just to make music, since (SPOILER ALERT!)…...I cannot find any detectable sonic advantage to any OS over another.


    Apple vs the Audiophile: other issues


    There are many myths and misunderstandings about audiophile quality playback on Macs.  I was amazed at the number of Google hits on queries like “high resolution audio MacOS”.  I was even more amazed at the variety of erroneous responses like “Macs are limited to 24/96” and “You can’t play FLAC files on a Mac”.  The bottom line is that you can play anything on a Mac that you can play on a PC or Linux box with the right settings and ancillaries – and you can do it just as well.  But you can’t do it all with iTunes or Apple Music – you’ll have to use 3rd party options, of which there are many.  


    The popular paid media players like JRMC and Roon work equally well on Mac, PC, and Linux devices of equal capability, and they’re very nice ways of managing your library and playing your music.  My “Value Proposition” review of front ends includes a number of Mac players that will do it all and do it well, and I only scratched the surface.  For example, BitPerfect which I did not include because it’s $10) is an excellent player that will let you listen to all your files except DSD in their native state on a Mac.  You can even play DSD, although it requires using their $30 DSD Master app to convert your DSD files to “hybrid DSD”(which, from its description, seems to be some form of DoP) – and the combined cost is about the same as a license for JRMC.  The BitPerfct website currently says that a second app called DSD Perfect is “coming soon”, but neither details nor prices are posted yet.


    If you want to add a sound card or interface to any current Mac but a tower, your options are almost all external.  Their 1/8” analog audio output jack used to be an optical S/PDIF jack as well on many models, so you could drive a DAC optically from it.  But Apple began removing this feature from select models in 2016 and I’m pretty sure no current Mac model has it.  So you have to use USB if you want an external audio interface, which leaves those whose DACs lack USB inputs in need of a new DAC.


    The bottom line on MacOS for audiophiles is that it’s as fine as any alternative, if you do it right.  You can get top quality sound from a Mac with no more work than it takes to do so from a PC, a Linux box, or any other platform.  So the decision point is, once again, how much you like the GUI, the work flows, the hardware, and (more importantly for Macs than for PCs or Linux) the source – if you love Apple, you’ll love MacOS.  But if you don’t like restrictions imposed on you by others, Apple’s even a bit more difficult than Microsoft in my experience.  


    Support is excellent, but when out of warranty or after AppleCare expires, it can be costly.  Apple is not very tolerant of those who don’t follow their party line, and they make unannounced and often unwanted changes that negatively affect the audiophile experience.  Removing the optical output is just one example.  Their juggling of iTunes, hi res, lossless, streaming etc is both annoying and sometimes very hard to deal with.


    As a final note on Mac, most of their hardware looks and feels great – it’s easy to love and makes a good impression on guests.  It’s more expensive than equivalent hardware for other platforms, and in my experience it’s no more or less durable.  Not all Macs have been so good (e.g. the 2015 12” MacBook was very poor), and they all suffer from some common problems like kernel panic and incomplete updates.


    None of these issues is a deal-breaker if you love Macs and want to use one for music.  They’re no more trouble prone than any other platform, and are generally quite reliable.  You’ll likely get years of listening pleasure if you simply set it up and enjoy it.

     


    LINUX


    THE MYTH, THE LEGEND, THE TRUTH


    Now that the easy stuff is out of the way, let’s dive into Linux. After reading to this point, you may well ask why you’d ever want to run Linux.  Whether or not you should even consider it brings us back to that prioritization effort I stress so often.  Like other operating systems (as described above in “the secret life of operating systems”), Linux is built around a kernel and requires many associated software packages to let you do what you need to do.  One of the main differences between Linux and WinMac from the user’s point of view is that the former lets you choose among alternatives for almost everything, while the latter makes almost all of the choices for you.  But one fact stands out -  it ain’t your father’s Linux any more!  The first Linux kernel had about 4500 lines of code, but current versions run as large as 23+ million lines .  For those of you who don’t understand coding, this is a VERY large program that approaches Windows  (which is probably now about 50 million lines  – I couldn’t find a confirmable number) in size and complexity.


    An operating system is a code bundle that includes the operating kernel plus all associated software needed to let users direct the computer to do what you want it to do.  Windows and MacOS are delivered as packaged operating systems with controls, presentation environments, communication tools, configuration systems etc already installed and ready to use.  You can change some of the basic appearance (fonts, colors etc), reskin some GUIs, and rearrange some screen architecture – but you’re largely confined to the functions and graphics built into the package chosen by Microsoft or Apple for you to use.  You pays your money, but you can’t take your choice!


    Unlike Windows and MacOS, Linux kernels are available with a dazzling variety of programs, user interfaces, system services, shell utilities, graphics, functions etc.  These associated and supporting programs are almost all open source and are provided (and maintained, in most cases) by volunteer developers in the open source community.  In the “good old days”, Linuxmade you choose a kernel and then find the supporting software you needed, package by package.  You then had to compile it all into a functional operating system yourself, which meant configuring and interfacing all those parts and making them work as an OS.  Many Linux people still do this, because you can specify and control every aspect of every interaction among all the software and hardware.  But it’s a world unto itself, and no sane audiophile would learn how to do this just to set up his or her stereo system.


    Fortunately, Linux-based operating systems are now most often delivered as bundled packages called “distros” (because each is distributed complete with a kernel and the necessary software to function to its specifications).  Distros vary from dead simple to fancy and sophisticated – but each is distributed as a working OS with interactions, dependencies, drivers, etc already compiled for you.  If you know what you want and need, and you pick a distro spec’ed to do it, you’ll have a system I’d compare to “made to measure” clothing.  It’s more like made-to-measure than bespoke.  It wasn’t designed and built from scratch to meet your needs - it was tailored to your needs from production parts modified as much as possible and necessary to do what you want. And it’s a pretty good system.


    Every operating system includes components on which other parts of it depend, e.g. runtime environments.  Many code packages are either dependent on others or are themselves dependencies for others.  Windows and Mac users are no strangers to programs that won’t run until another program like JRE or Flash is installed.  But Linux seems to have more such relationships than WinMac, as installation of new software triggers a fair number of “unmet dependency” warnings.  In all fairness, most of the best distros for audiophiles include package management programs that make installation of new apps as easy and painless is it is on Windows and Mac.  But if you choose to install an app like player from the command line (which may be necessary if it’s not available from the “software store” associated with your distro), you can spend a lot of time finding and installing those dependencies


    There are dozens of graphic & functional environments available for Linux – here’s a guide to the most popular ones., and with a little work you can actually create your own.  Many of the most popular and easy to use distros offer a choice of desktop and graphic environments on installation, and not a few distros differ from each other primarily in the choice of the GUI(s) included.  When last I looked, Ubuntu 18 provided the Mate environment but 19 moved back to GNOME. Fedora includes GNOME, and Mint offers Cinnamon.  But you can easily (OK – it’s not always easy) install an alternative desktop / graphics environment once you learn how.


    Linux is not for everyone, and it requires you to make some compromises that you may resist.  Many flavors also call for knowledge, skill, or interests you don’t have and may not want to develop.  So I hope this section will be a helpful guide that will let you match Linux distros to your value scale.


    Almost anyone can learn to use Linux if the desire is there – it ain’t rocket science, and it’s quite logical. I’m amazed at the number of people who tried it only because it’s free, and subsequently fell in love with it.  Once you learn the syntax and how to communicate with it, you can easily work your way through many distros, adding functionality and making a really nice audio system of which you’ll be very proud.


    I strongly recommend cleaning up and renewing a computer before resuscitation with Linux. Add a new SSD for the boot drive & get RAM up to 4+G, although 2’s fine for lighter distros if that’s max.

     


    WHY LINUX?


    Former Windows and Mac people adopt Linux for many reasons, including

     

    • legacy operating systems that have reached the end of life
    • need or desire to refresh and maintain legacy hardware but upgrade OS
    • hardware failure (e.g. HDD) & loss of OS with no backup or restoration available
    • ability to extend computing capabilities on existing hardware limited by Win/Mac req’s
    • lack of desired flexibility in Windows or Mac
    • forced conformance to MS and Apple practices, policies, and restrictions
    • ability to update and upgrade OS ad lib at no cost
    • ability to try, run, or switch among multiple operating systems with no restrictions
      • yes, Win10 has an “Ubuntu” option
        • but it runs in Windows Subsystem for Linux (WSL)
        • no, this is not a Linux kernel – it has no utility for audiophiles that I can identify
    • a love of choice
    • a pioneering spirit


    And there are many advantages of Linux over WinMac, including

     

    • Linux is bootable “live” from a USB drive
      • you can try as many distros as you want without altering your current computer setup at all
        • burn each to a USB stick, plug it it, make sure your BIOS or UEFI is set to boot first from USB, and you’re off & running
        • make sure you understand the limitations on what you’re using; some will store configuration data in the USB drive, while others will be back at the starting gate again with each boot
    • many Linux distros offer the option to install as a dual boot alongside your resident OS
      • much better security than Windows and better than Mac
      • easier administrative control over user access to programs, files, directories etc
      • root access is not granted by default to any programs in Linux
      • there’s no registry in which to bury malware and back doors
      • despite use of Linux on many (if not most) of the world’s web servers, attacks are rare compared to Windows environments
      • updates and upgrades are free and easy
      • apps are much less frequently rendered useless by upgrades, compared to Windows
      • a large and active community of Linux users, developers, and enthusiasts who provide unofficial but often very helpful support through wikis and forums
      • community response time is most often same day and usually much better that the unpaid “support” offered by Microsoft and Apple through web forum participants
      • responses and proposed solutions to problems differ, just as they do on all other forums
        • but feedback is strong and quick, so it’s often easy to figure out which solutions worked for most people and which were not helpful
    • current distros are USB compliant and work without separate drivers for Linux-compatible stuff 
      • this is especially important for audiophiles, whose digital system components may not all be compatible with Linux (e.g. DACs and DAIs)
        • be cautious and expect to have problems with some stuff even if new and current
          • I had to abandon several pieces of my recording setup when I went from Win7 to Ubuntu a few years ago because they were simply invisible to my computer, e.g. TASCAM US-1800 8x8 DAI
          • even today, things that should work and are stated to do so by the source do not always work; I bought an inexpensive 7.1 USB DAC a few months ago because the manufacturer’s website said clearly and repeatedly that it worked fine with Windows, Mac, and Linux.  It didn’t.  I returned it and bought a Xonar sound card that clearly said it worked with Linux.  It didn’t.

     


    WHAT’S TO CHOOSE FROM?


    As I said above, there are well over 100 free distros, adaptations, derivatives etc available and in common use right now.  Some are light and easy on resources, which lets you do more on older or more basic computers than you can with the OEM system that came on it (even if updated to the limits of the hardware).  Some distros rival Windows and MacOS for their complexity and ease of use, with very nice GUIs for configuration and excellent software management packages.  I know of no Linux package that’s as bloated with code as Windows, but some of the major distributions are far from lean.


    Others present themselves as “Windows-like” but are less cohesive, having straightforward desktops and GUIs for common tasks but occasionally requiring you to find and use an oddball tool or enter command prompts in a terminal window.  And many are quite spare in out-of-the-box content, requiring you to set up even basic functions like networking and file sharing that you never even thought about when running Windows or MacOS.


    Almost all of the Linux packages you’d consider for audio are open source and available free of charge, although there are some weird restrictions on a few, such as Red Hat’s prohibition against copying the code verbatim. A fairly robust industry has grown around paid support for Linux that’s little or no different from free, open source versions.  For example, RedHat is an established, large commercial enterprise that provides “... production-ready code, life-cycle management, software interoperability, and access to experts and tools to help you run your business”.  Their menu of products and services is impressive, and they will customize Linux instances as you wish for a price.  But you can download and use Fedora or CentOS, which are “community editions” of the same Red Hat Linux OS you’d get if you bought it from them.  


    You don’t need commercial support to make great music – there are a lot of excellent distros available at no cost, many of which (e.g. Ubuntu) are very easy to install, configure, and use.  I strongly doubt that an audiophile would want or need to pay for Red Hat Enterprise Linux (RHEL) solely for music management and delivery.

     


    USER INTERFACES


    It helps to understand the concept of a desktop environment, which is the tool set you use to make your computer do what you want it to do.  This list of many of the basic components may help you visualize:

     

    • Browser: Browse the internet.
    • Command line access, e.g. terminal
    • Desktop communications, e.g. widgets, panels
    • Desktop controls, e.g. menus, launchers, system tray
    • Display Manager to control the graphic parameters of your GUIs
    • File Manager
    • Productivity basics, e.g. office suite, text editor
    • Window manager: Managing themes and how windows behave.


    Microsoft and Apple offer their own distinctive graphic environments for your desktop, home screen, backgrounds, controls etc.  From day 1, you know how to apply the limited customization they allow, how your working screens will look, where to find the icons and tools you need, how to open menus etc.  Updates and upgrades (whether mandatory or elective) may bring changes in these things – and if they do, you’re stuck with most of them whether you like them or not.  Windows has preserved some legacy screen layouts for those who didn’t want to adapt to the “new improved” ones – but such flexibility is less common with each new upgrade.


    Linux is a whole different world in this regard.  Although most distros come packaged with a specific graphics environment, and many users pick the distro on this basis, there are many, many alternatives that will work with most distros.  You simply have to download, install and configure them (which can be easy as pie from a box or hard as pie from scratch, depending on your distro and choice of environment).  Unfortunately, many Linux applications are written for a specific environment and will not work unless the necessary elements are installed as well. So if you decide you want a different graphic environment, you may find that your desktop and/or installed apps will no longer work properly or at all.


    If you just want a Linux-based music system that you can enjoy with minimal setup, configuration and ongoing intervention, pick a popular and well supported distro with a user environment that you like, and stick with it.  Only install software from the PPAs associated with it or use proven programs with strong online support, like Roon, JRMC etc.  When you have it working well, leave it alone!

     


    ACCESSING YOUR MUSIC FILES (NETWORK FILE SHARING)


    Linux can be a bit confusing when it comes to file sharing with Windows and Mac devices.  Although you can readily share files among all your devices, some distros (e.g. Mint and Ubuntu) make it easier than others because they’re installed with the necessary packages.  There are a few commonly used communication protocols for file sharing, including

     

    • Server Message Block (SMB)
      • local network communication tool for shared access to files, printers, and ports
      • created at IBM almost 40 years ago and upgraded multiple times since
      • Common Internt File Sysytem (CIFS) is an early version of SMB created by Microsoft
    • Network File System (NFS) created by Sun; useful only for all-Linux networks
    • Apple File Protocol (AFP) is an older Apple creation formerly known as AppleTalk
      • Macs now use SMB over TCP/IP but not NetBEUI (a Microsoft network transfer protocol)


    The industry standard for sharing files between Linux and other computers has been an open source package called Samba for almost 30 years.  Samba is a collection of software tools that transparently manages file serving, authentication, name resolution, and print services.  It’s included in the installation packages of many of the better distros, and it’s easily installed in the rest.  With Samba installed and properly configured (which is also easy), files on all networked devices will appear in all file system managers on all networked devices (e.g.Explorer in Windows and Nautilus in Linux).


    It can be a bit of a bother to make Samba work in some distros as we’d like it to do.  For example, less robust Linux versions make you alter a configuration file from the command line in order to remount your network file shares on reboot.  If your music files are on a network server, you’ll have to make sure you have a permanent mount of the relevant directories if your distro won’t maintain it, or your library will be empty every time you open your player without remounting after every reboot.


    The best distros will be an integral network participant when installed, and you will not need or have to do anything beyond opening your directories in an included file manager that’s virtually identical in appearance and use to Windows Explorer.  You’ll be prompted for access credentials (if necessary) and asked if you want them to persist so you don’t have to redo it every time.  Some Linux distros are also sticky about directory and file access permissions, and the occasional “You do not have permission to access this file” alert will pop up in many installations.  I have not had this happen at all with Mint or Ubuntu 18+, and it’s not been a problem at all with any of the Raspberry Pi operating systems I’ve run. But it’s a real annoyance when it happens, so you need to be sure to use the same user account to install and set up all your file shares and to add files to your library, to avoid problems with file ownership. If you create files while signed in as root, you have to assign permissions to other users or the dreaded access denial will thwart your listening pleasure.

     


    LINUX SOFTWARE MANAGEMENT


    Because Linux is entirely (for all practical purposes) open source, it’s a conglomeration of inputs from thousands of developers and teams with varying levels of sophistication.  The best distros are rock solid and (in my opinion) rival or exceed Windows and MacOS in stability, capability, and user friendliness.  But there are many that make you accept one or more compromises for some very sophisticated functionality.  If you love something about a weird distro and you can’t get it in standard packages, you may have to live with bugs, instability, and/or frequent (even daily) updates and changes.


    Linux comes in several levels of stability.  The stable, long term support releases (usually shown as LTS after the version numbers) are probably best for most audiophiles because they’re stable and will receive support (updates etc) for a decade from release. All now require PAE support in the hardware.


    Updates can be a hassle, but they’re generally not as potentially irritating or disruptive in Linux as they are in Windows.  I think it’s easier to configure the good Linux software managers (e.g. the one that comes in Ubuntu) to alert you to the availability of updates and let you choose the ones you want to download and install than it is to do this in Windows (which somehow seems to know the exact worst moment for installing all available updates at once and rebooting). 


    The update process for Linux is a bit different than that for Windows and Mac.  Because all Linux software comes from PPAs, your computer has to check the PPAs associated with your distro and any 3rd party sources you added yourself every time.  It generates and keeps a list of available updates for every installed program every time it does this, and displays a list of all programs with available updates so you can select the one(s) you want to install.  I only install those that affect my use of the device, to avoid the occasional unpleasantry an errant update can stimulate.  For audio use, that means updating the operating system and environment, any and all apps associated with the process of sourcing / playing / indexing / managing music files, their dependencies, and any networking software.


    You do not have to update anything after installation if it’s only a dedicated music player.  Left entirely to its own, any good distro will probably be stable for a very long time if only playing music.

     


    THE DISTROS I RECOMMEND & WHY


    I’m a big fan of Linux Mint for many reasons.  It’s based on Debian and Ubuntu, so it has a stable, proven code base and about 30,000 software packages available for it (along with excellent software and update management programs with fine GUIs).  It runs very well on older machines with minimal resources.  Although installation can be slow on them (see the OS comparison chart for details), I’ve now put it on half a dozen machines from the dawn of the 64 bit age (2005-6) without a hitch, and there’s still a 32 bit version that I’ve used on a few old machines with fine results.  I just ran a fresh installation of the latest 19.2 version with the Cinnamon desktop environment, and it’s better than ever.  


    I’ll run you through the installation and audio setup process so you’ll get an idea of the work involved.   Although it’s a lot faster on fresh machines with modern resources, it goes as well onto any machine that will run it – the only difference is the time it takes to complete the installation, setup, and configuration.  So this example is from a late 2005 Toshiba Satellite U205 with an Intel T5200 2 core CPU running at 1.6GHz) and 2G of RAM.  Boot to login now takes 17 seconds and general response time is on a par with my Win10 machines.


    As for most Linux installations, you just download the image and burn it onto a USB stick.  You’ll be offered a choice of desktop environment with Mint and many other distros when you choose your download (the one below is called Cinnamon – Mint offers you 3 choices with 19.2).   Make sure your BIOS is set to boot from USB, insert the stick into the machine you want to set up, and power it on.  With Mint and most other distros suited for audiophiles, you’ll get an option to try it from USB or install it (the circled icon below on the left is the installation link).  If you run Mint from the USB stick image, you’ll learn how it looks and feels but will not be able to save any data.  Some distros can be used from USB with data persistence, but it’s unusual in distros for PCs. The right screen below is the installation choice selector.  From this, you can install as a dual boot or overwrite the existing OS (which is permanent, so work carefully!).  You can also encrypt the installation if you wish.  


    mint_install_screen.jpg mint_install_choices.jpg


      

    During installation, you’ll get a series of choices that will make your instance look, feel, and work as you want it to do.  On completion, there’s a reboot prompt that will warn you to remove the installation medium. And once it boots from your HD, you’ll get a series of prompts to finish your setup so you can get to work.  The left screen below is the GUI for setting up a lot of management tools easily, and the right one is the software management GUI for Mint 19.02 -
      

    mint_mgt_tools.jpg mint_software_mgr.jpg

     

     


      

    Remember that you can use whatever static backgrounds & desktops you want in a Linux distro – photos, graphics, etc.  There are also software packages that let you create and use dynamic backgrounds, including both videos and active functions like stackable icons.  The graphics environment only determines how the control elements and forms will look, e.g. menus.   You can change color schemes as well, and there’s a wealth of alternative graphics packages out there for the hard to please.  This work addresses as delivered distros only – most players have their own skinnable graphics packages.


    Here are two common Linux desktops.  The left is the default Cinnamon environment on Mint 19.2, and the right one is an Xfce lightweight desktop on Ubuntu Studio (with the software updater alert).  I ran this instance of UStudio for several months for live audio recording and mixing, before dropping Mint on the machine to document this article.
      

    mint_cinnamon_desktop.jpg xfce_on_ubuntu.jpg

     

     


      

    Like Mate, LXQt, and others, the Xfce environment (on the right, above) uses far fewer resources than fancier milieux like Cinnamon & GNOME.  This makes it faster on older and/or less powerful computers - but you give up features for that sleekness.  The LXDE environment / desktop is probably the lightest of all – it launches in seconds on any machine and uses only about 300 meg of RAM.  It’s so small that it’s the environment in Raspbian (the dedicated OS for Raspberry Pi).


    To be honest, I don’t care much which desktop I have on a computer that I’m using primarily for music. In my experience, most audiophiles don’t waste much time on graphics they use only to boot and power down – it’s the music programs that really matter.  So I try to minimize demands on CPU, memory, and data transfer routes throughout the computer, to free up resources for audio.  Nevertheless, I like Cinnamon for its flexibility, speed, and clean feel.  I use it on the Gateway media center that houses my DAW (Ardour) and audio/video production & editing tools (Lightworks, a powerful tool available for Win, Mac and Linux in 32 and 64 bit versions).  But for older computers, especially 32 bit machines with 2G or less of RAM, I’d go as light as possible.


    Ubuntu Studio is self described as a free and open operating system for creative people – and it is!  It contains software for audio, video, graphics, photography, and publishing – and all of the included programs are good enough for pro work, although they’re not sufficiently robust for professional production work flows. True pro products go far beyond these, but they also cost a lot more and are suited for large scale production work.  I wouldn’t be surprised at all to learn that Oscars and Grammys have been won with material created on Ubuntu Studio.   It’s fairly easy to install and configure, although it includes some audio tools (e.g. Jack) that can confuse the novice and make setting up your audio player and playback paths a bit more complex.  Ubuntu Studio requires a robust computer with 


    It includes many packages for audio, but its focus is audio production – recording, editing, mastering, etc – which is why it uses a low latency kernel that reduces delay to <5 msec.  In simple terms, this is achieved by prioritizing audio calls to the CPU over requests from other sources.  This requires frequent queuing to maintain order and priority among the programs banging on the processor’s door, which reduces throughput in favor of rapid audio processing.  If you’re running a current multicore multithreading CPU, a low latency kernel is probably superfluous for nonprofessional recording.  Latency is the bane of low budget performers like me who build a recording by laying down multiple tracks one by one, because even a tiny delay is uncommonly distracting when the monitored live track is not perfectly time-aligned with the playback.  I’ve been plagued by it for years – but it’s gotten better with each advance in processing power, and many DAWs have built-in corrections you can adjust to time align the monitoring of your live signal with the playback of other tracks on the recording.  


    None of this matters at all to audiophile playback.  I hear no difference between the same players on Ubuntu Studio and Linux Mint driving the same DACs and audio systems.  If you do, a low latency kernel is easy enough to install for almost any distro – it only takes a few commands in the terminal window and (depending on the distro) a little configuration.  But the easiest way to get it if you want it is to simply use Ubuntu Studio.  System requirements start with 10G of HD space and 1G of RAM, but Studio will not run well for you with less than 4G of RAM and 20+ G of HD of available HD space.


    Ubuntu Mate is my pick for older, lighter computers – it’ll run well on just about anything you can boot.  Mate is a lightweight desktop environment that’s a current development of the older GNOME 2 environment used by Unix and Linux lovers back in the day.  This distro includes some nice apps you’ll probably like and use, e.g. Firefox, LibreOffice, and Rhythmbox – but it’s not bloated with stuff you never heard of and don’t want or need.   


    Mate available is 32 and 64 bit versions, and it’s a great way to breathe new life into an old box.  It’s also available for Raspberry Pi and the GPD MicroPC (a cool little device that gets no attention in the audiophile press).  The GPD looks to me like it has great potential as a central controller for a LAN-wide audio system.  If I read the spces correctly, it should be powerful enough to run JRMC or Roon on either Win10 Pro ( its default OEM OS) or Mate (also available through and supported by the manufacturer). I gotta get me one of these to play with!


    Peppermint is a relatively new entry to the world of Linux, but it’s based on the extremely stable Ubuntu 18.04 LTS kernel and is designed to be light, easy to use, and “web centric”.  It’s kind of a code mongrel, drawing what its developers believe is the best of each function from currently available top quality Linux distros.  The default desktop environment is a version of the lightweight Xfce4 I showed above in my current Ubuntu Studio instance.


    Instead of bloat in the installed package, Peppermint offers options - their stated philosophy is that “[w]e think the user should decide”, which (at least in my opinion) is a great approach. You can choose from among multiple alternatives, starting with the browser to be installed.   Choose from Chrome, Chromium, Firefox, and Vivaldi directly from the installation menu (which is a simple, easy to use GUI).  The software package manager was developed in house at Linux Mint (and is installed with Mint as well).  With a very nice, simple GUI, it makes finding, installing, modifying, and removing software as easy as clicking “install” etc.  It’s at least as easy as the Windows approach, if not easier.


    One major goal of PeppermintOS is to make local apps and web apps equally accessible, usable, and enjoyable. So you can opt for a minimal installation with only the basic OS and a web browser, which results in a functionality similar to Chromebooks without the tight Google-based logins and linkages. It facilitates reliance on web and cloud based data sources and is in some ways a reincarnation of the old dumb terminal – mainframe architecture.


    I thought this might be a good approach for audiophiles who favor web streaming, so I installed and set up a Peppermint instance on my old Toshiba Satellite (one of the first 64 bit machines, with only 2G of RAM).  Installation is very similar to Ubuntu’s, for obvious reasons – it’s a proven, simple and stable process, and the kernel driving all this is based on the 18.04 long term support code of Ubuntu.  The entire process from downloading the image to burning a USB stick to playing music took about half an hour on this older machine, including automatic downloading and installing of 15 package updates.  Once set up, boot takes about 15 seconds.


    Peppermint does install with one media player called “Media Player” that strongly resembles (and may in fact be based on) VLC, although its properties and setup are simpler than VLC.  It plays FLACs out of the box, but does not recognize DSD files.  And it does MC up to 7.1, although I didn’t test that out.


    Like most Linux distros, Peppermint does not install with a full complement of utilities – you have to add one to be able to download and set up Roon Bridge.  The instructions for doing this are clear, simple, and provided on the Roon downloads page.  Yes, you have to enter 3 commands into a terminal window one at a time – but this is very easy and shouldn’t dissuade anyone who can set up and operate an audio system.  Downloading Roon Bridge requires a data transfer package called cURL (an internet protocol for file transfers).  There are alternatives to curl (e.g. wget) used by other software sources, so you need to read the instructions for downloading and installing players etc to be sure you have the right package(s) in your distro.  The Roon Bridge installation took less than a minute from starting the process to playing music.  


    I then loaded a few web radio players, the coolest of which turns out to be Goodvibes.  This is a current  package that uses the latest Gstreamer to manage file formats – so it should be able to stream high resolution audio to any hardware that can play it.  The current Gstreamer will do DSD, if I read the documentation correctly, but I couldn’t find any streaming DSD to try it.  I also tried Radio Tray, a very light web radio streamer that’s readily configurable and easy to use. And for good measure, I put the Spotify Linux client on the same machine by following these instructions from Spotify, and it works great. 


    Browser based streaming is as easy as opening your browser and entering the URL for the source you want.  Peppermint includes a very nice feature, excellent for browser based web streaming, called “ICE”.  Accessed from the Internet section of the app menu, it enables the creation of site specific browser windows (SSBs) in the browser of your choice (Chrome, Chromium, Firefox, or Vivaldi) for specific URLs.  These SSB windows are almost entirely “chromeless”, i.e. they have few of the usual forms and controls found in a browser window.  


    Each then appears in the app menu that’s accessed from the lower left menu icon, with either the site favicon or an image of your choice next to the name you gave it. So you can open your browser-based web streaming sources just as you would a player or other app, simply by clicking the icon in the menu.  Here’s what it looks like on my Toshiba – I just click the Amazon Music icon (white arrow added to picture for clarity) and up pops the web player in a browser window.

     

    peppermint_ice_amazon.jpg

     


      

    ADDING AN AUDIOPHILE PLAYER / MUSIC MANAGEMENT SYSTEM


    Many Linux distros come with at least one rudimentary music player in the package, and there are others available in the PPA associated with it.  Mint is installed with both Rhythmbox and VLC players, so you can have your music up and running within minutes of completing the installation if you’re happy with either of them.  Installing JRMC or a Roon package is not difficult, although you have to use the terminal window and enter commands one at a time to add the repository in which the files are located, update your computer’s package inventory to add what’s available in the new PPA and install it.  


    On the left below is the GUI view of a simple Roon Bridge installation on my Ubuntu Studio Gateway PC.  The instructions on the Roon website are clear and simple.  Although those not used to working with a command line or terminal window may fear it, you can almost certainly master Linux basics with no difficulty at all if you can put a component stereo system together and keep it running well.  You can learn what’s entailed by reading the instructions for JRMC and Roon installations on Linux.


    To install either Roon server or Roon Bridge on a Linux machine, you simply have to open a terminal window and enter 3 commands, one at a time.  The left image is the installation terminal open over the Roon instructions, and the right image is the Roon audio settings panel after installation showing the addition of the Toshiba to Roon outputs.  The entire installation took 2 minutes from accessing the Roon download / instructions site to playing a tune.


      

    roon_bridge_install.jpg roon_audio_config.jpg

     


      

    FURTHER THOUGHTS ABOUT LINUX DISTROS FOR AUDIOPHILES


    All current Linux distros are USB 2 compliant, so you don’t need drivers for USB audio data transfer to your DAC (except for the occasional hardware that doesn’t play nicely with Linux, which is fortunately becoming rare).  Some Linux distros (e.g. Ubuntu) began supporting USB 3 in 2009, and AFAIK all of the ones I’m discussing here are fully USB 3 compliant..Your computer has to have a USB 3 port in order to use that capability. There’s an easy way to find out if yours has USB 3:  just open a terminal window and run the command lsusb to display all your USB resources.


    There are several excellent distros with interchangeable capabilities for audiophile use.  Here’s a summary table of those I consider to be among the best for audiophiles.  With well over a hundred stable, proven alternatives out there for us, I don’t claim to have covered the field and I’m sure there are several that would please me as much as my current choices in actual use.  But remember that SQ is going to be the same for almost all of these, assuming they’re configured for bit perfect, lossless playback at the resolution of your choice.


    One major find for me in this project was Peppermint, with which I was totally unfamiliar until I started researching this in depth.  If I were a betting man, I’d wager a fair amount that there are several more great ones awaiting our discovery.  Many of you probably know of at least one, so feel free to suggest more in your comments.

     


    Audiophile Linux


      APLinux.jpgI decided to include Audiophile Linux because so many audiophiles ask me about it.  Yes, it’s a Linux distro, and it does a fine job on audio playback once you get it installed and configured (which is a bit tedious – see the link in the summary chart for a full description of the effort required).  It’s not an ideal or easy platform on which to build a computer for any purpose other than music playback.  The effort required probably outweighs any benefits and potentially degrades its use for audio.  It installs without the productivity and utility packages you’d need to do anything else and includes only a light desktop environment called Fluxbox that you’ve probably never seen before.  Yes, you can use the Arch package manager to install almost anything you’d want – but adding much other stuff turns it into just another distro that happens to have a real time kernel.


    The installed audio back end in APL is good old MPD, and installation includes the very nice MPD client Cantata by default.  Say what you will about the age and simplicity of MPD – it’s still an excellent player with high resolution capability.  And Cantata is a very nice front end client for MPD that offers pretty good library function. If all you want to do is play music, I can’t imagine a much better no cost solution as long as you’re willing to jump through the installation and configuration hoops it includes.


    The Arch package manager (called pacman) lets you access over 50,000 programs including players like Audacious, DeaDBeef etc.  So you’re not stranded on a desert island with APLinux - but you may feel like you’re trapped in the middle of cyber-nowhere if you want to make the box behave more like a “normal” computer.  So APLinux is truly for audiophiles and absolutely not a good general purpose Linux distro.  It currently installs with Chromium and Firefox as browser alternatives


    It installs with a real time kernel, which means essentially zero latency (well under 5 msec) in executing audio calls.  But as I’ve said before, I’m not at all convinced this matters for even the highest quality playback unless you’re using an old &/or low functioning CPU with few cores and limited hyperthreading ability.  No - “it can’t hurt”, and it’s absolutely necessary for live recording. But that’s simply not true for audiophile listening from a robust machine with few background processes running.


    NB: Audiophile Linux is NOT Audiolinux - they are two completely separate products.  Audiophile Linux is a freeware distro based on Arch Linux with a real time kernel and an MPD back end – so it qualifies for this comparison article because it’s a practical open source alternative to other Linux distros.  Audiolinux is a paid distro, also based on Arch Linux, but with more advanced features like the ability to run the entire system from RAM  (“ramroot”, which requires at least 16G with a GUI and 8G headless).  It’s a niche commercial product that does not compete directly with JRMC, Roon etc because it’s not a player or management system – it’s purely server side software.  So for purposes of this comparison, it competes with free Linux distros by charging users for its claimed superiority to no cost alternatives.  You still need a player / endpoint, for which you can use anything you like – Roon, JRMC, hardware network players / streamers etc.


    Audiolinux more appropriately compares to other Linux based music systems like Euphony.   Unlike Audiolinux, Euphony has its own player and management system and can be used as an OS with other software.  I think it now comes with an embedded Roon server (requiring a Roon subscription if you want to use it) and other endpoint software.


    I have not purchased it and I don’t know anyone who has - so I have not seen, heard or used it.  All I can say is that it’s another low (OK, very very low) latency Linux distro that should be a good choice for you if it does what you want better or more completely than the alternatives.  But if I had to choose between buying Audiolinux and using a freeware player or buying JRMC and running it on Mint or Ubuntu, I’d go for JRMC and optimize my computer as well as possible.  Some reviews suggest that the author’s player of choice sounds better than the same player on any other distro – but I have yet to see a proper comparison of Audiolinux with another distro like Ubuntu Studio.


    I’d love to discover that Audiolinux sounds better than the alternatives – I’m just a bit skeptical.  Ramroot may improve SQ, but it’s hard to imagine that anyone would put 16+G of RAM in a marginal computer – and once you build beyond a certain point, SQ differences are vanishingly small even without running from RAM.  


    Here’s a short summary of some observations of Audiolinux and Euphony posted on AS.  I think it’s illustrative of the difficulty of judging these things for yourself because 1) they’re truly very close, 2) the variability from listener to listener is high, and 3) the drama with which differences are reported seems a bit exaggerated to me:


    “Running AL the ring of the inital [sic] Gong hits are almost recessed.  They just do not sound right. Reboot into Euphony and that wonderful ring is there and the decay just lasts forever.”


    “Audiolinux (set up Extreme, Ramboot [sic – is this supposed to be ramRoot?]) sounds comparitively [sic] flat and muted, as you say the decay and representation of the instruments is better on Euphony - it is an easily dicernable [sic] difference.”


    “I have personally tried both Euphony and AudioLinux and found without doubt that Euphony is better sounding.”


    “I compared my install of AL with the demo version of Euphony on a USB drive. At that time I found that AL was quite a bit better than Euphony.”
    “If I may use analogy, AL vs. Euphony resembles me HQP vs. Audirvana. “


    “I find that AL in both machines sounds lean. I have to adjust my DSP and subwoofer settings to have higher bass than my usual room curve to give me the much needed bloom and meat. With Euphony on both machines, I set to my normal room curve but I miss the. Transparency and PRAT of AL [sic]. It sounds quite boring.”


    “I tried also playing into RAM. I never liked this option in AL, no difference in Euphony. Hard, harsh and do I sometimes hear distortion in fortissimo string passages.”


    I don’t know what to make of descriptions like “They just do not sound right”,  “It sounds quite boring”, and  “...much needed bloom and meat”.  When one of the reviewers who wrote these statements said “I found that AL was quite a bit better than Euphony” and another said “representation of the instruments is better on Euphony - it is an easily [discernible] difference”.


    I’ve stuck to open source, free Linux distros to avoid the question of whether you get better SQ (or anything else) if you buy one of the distros aimed at audiophiles than you do from any of those that are widely available at no cost.  So I’m not going to address paid distros like these at all because they’re all so close in SQ that as far as I’m concerned) you pay your money and you take your choice.  

     


    SUMMARY CHART OF MY FAVORITE LINUX DISTROS FOR AUDIOPHILES

    SUMMARY CHART OF MY FAVORITE LINUX DISTROS FOR AUDIOPHILES.jpg


        

    SUMMARY THOUGHTS


    There is probably no significant, consistent difference in SQ among operating systems that have been set up and configured properly for audiophile use on computers that meet system requirements.  The variety of impressions about SQ among equally experienced audiophiles is staggering over the spectrum of operating systems and hardware, and there is absolutely no consistency over time or across populations.  Putting this into the context of the above quotes about the SQ of Audiolinux and Euphony, I truly believe that the top operating systems (open source or paid commercial) are too close to separate when installed on capable boxes and optimally configured, adjusted, and tweaked for audio playback.  I therefore find no reason (for me, at least) to purchase an operating system when I can get top quality sound from open source software, and I stand by that recommendation for most people.


    If you’re going to base your system on a Windows or Mac platform simply because you like the OS or because you’re limited to an existing computer for any reason, you can also get top quality sound from either one.  We’ll look at x86 vs SBCs etc in a future installment – but if you plan to run your music server and player on the same machine, an x86 box is probably best regardless of your chosen OS.


    In the next installment, we’ll look at the little guys – SBCs, microcomputers, sticks, etc.  This is a world of JEOS (just enough operating systems), lean coding, high efficiency, and low cost.  You can do an amazing amount with these devices and play music with SQ that (to my ears) equals what you can get from a big old x86 box.  Stay tuned, especially if you love your current computers, have or plan to use a network server and storage, and just want to have networked endpoints on which to play your music.  An SBC can’t do it all, but what it does it does very well.


    You can have your Pi and eat it too!




    User Feedback

    Recommended Comments



    22 hours ago, bluesman said:

    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?

     

    Pretty straightforward ... how close is the SQ to the point where you are no longer conscious of the 'machinery' in your room doing its thing, and where the captured musical event dominates your awareness of what's going on, soundwise. 100% means that you've got that happening ...it can get better, because this level of subjective presentation requires the brain to compensate, unconsciously, for misdemeanours that stop it being 100% accurate in a technical sense - better means the brain needs to do this even less.

     

    Yes, the problem if you are sensitive to these aspects is that conventional quality ambitious sound is highly irritating - you become aware of what it's doing wrong, and you "can't stand it!". And, yes, the answer sometimes is that you give away the game of high end sound entirely; this is something I did for about 10 years, some time ago - you drop back to mid fi listening, because then you are no longer concerned about SQ; it's "kitchen radio" listening.

     

    Share this comment


    Link to comment
    Share on other sites

    3 hours ago, audiobomber said:

    You don't need to know much of anything about operating systems to play music through a computer.

    Right you are - as I said in the last piece in this series, "A Raspberry Pi 4 with 4G of RAM is an amazingly powerful little box that will run some great players from simple MPD to a full instance of JRMC - for $55 plus a power supply.  SQ is excellent, and it takes less than an hour from the time you walk in your front door with it to be playing music."  The piece before that one has a simple table with info offered to help uncertain audiophiles choose a player for the Pi.  It includes both player software running on named Linux distros and players that come with dedicated JEOS like your Volumio instance (which I reviewed, included, and use myself).

     

    But even for those who, like you, are using simple little boxes that require little active management, knowledge is power.  I'm in your camp - I'm currently running music software of some kind (including Audacity and Ardour for live recording) on two Pi 3bs, two Pi 3b+s, two 4 gig Pi 4s, and a Zero W.  The 3B+ can't hold a performance candle to the Pi 4 (which I document in the next piece, which is on SBCs and includes links to performance testing of the Pi family). So you may someday find yourself wanting to upgrade, in which case you'll want to know something about software related problems in the 4 gig Pi 4.  Start with the fact that it won't access more than 3 of the 4 gigs of RAM right now because of software bugs (at least on Raspbian).  And this affects USB connectivity, which came as a huge surprise to me and does not affect 1 and 2 gig Pi 4s.  The workaround is to limit RAM use to 3k in one of the configuration files - but I want the RAM I paid for 😉 

     

    There are a few patches out there to try to fix this, but it remains a problem and I have it - Roon Bridge drops and reconnects it as a zone in a cycle lasting roughly a minute per drop & add, and it appears to be because the USB connection to the DAC is being lost.  I don't know yet if this bug affects the 4 regardless of OS, so I'm testing it with several to find out.  So as long as your Volumio instance runs fine on your 3b+, you don't need to know anything else about your operating system.

     

    Remember too that your use of computers for audio is only one of many possible roles they can serve.  If your needs or preferences suggest that an x86 or other big box is the best home for your computer solution (or you want to use an older computer you already have but can no longer run its legacy OS), some kind of guide or information for decision support may be very helpful in choosing your new music platform - and that's why I'm writing these.

    Share this comment


    Link to comment
    Share on other sites

    15 hours ago, bluesman said:

    Right you are - as I said in the last piece in this series, "A Raspberry Pi 4 with 4G of RAM is an amazingly powerful little box that will run some great players from simple MPD to a full instance of JRMC - for $55 plus a power supply.  SQ is excellent, and it takes less than an hour from the time you walk in your front door with it to be playing music."  

     

     The 3B+ can't hold a performance candle to the Pi 4 (which I document in the next piece, which is on SBCs and includes links to performance testing of the Pi family). So you may someday find yourself wanting to upgrade, in which case you'll want to know something about software related problems in the 4 gig Pi 4.  Start with the fact that it won't access more than 3 of the 4 gigs of RAM right now because of software bugs (at least on Raspbian).  And this affects USB connectivity, which came as a huge surprise to me and does not affect 1 and 2 gig Pi 4s.  The workaround is to limit RAM use to 3k in one of the configuration files - but I want the RAM I paid for 😉 

     

     

    I have a Pi 3 B+. I would not describe the sound as "excellent", more like "good", as long as your DAC is separately powered from the terrible Pi USB port. When separately powered, the 3 B+ sounds better than CCA but I don't think there's anything you could do to a Pi to match the SQ of an sms-200.

     

    I had Pi 4 in my sights until I read this Hans B review. I've also heard concerns about overheating.

     

    Share this comment


    Link to comment
    Share on other sites

    You’ll read more about temp control & overheating in the next chapter.  And there’s more on this in the following one (being prepared now), which is about hot rodding SBCs for audio with a focus on the RPi. Cooling is essential if you overclock, but a good fanless case like the Flirc (my favorite) is all you need for a stock Pi 4.  Placed in an area with good ventilation at an ambient temp below 25C (about 80F), even a stock plastic case should be fine for most audiophile use of a stock 4. I strongly recommend using heat sinks though - they’re tiny, cheap, and effective. The life of Pi is shortened by running often at temps close to the throttling threshold.
     

    The stock 4 throttles CPU speed down to that of your 3b+ when its temp hits 80C.  I’ve pushed one hard with live multitrack recording and real time monitoring, and it peaked at 73C in the standard plastic case. A Flirc brings that down to about 60C, and a $7 fan cooled case lowers it to a max of 43C no matter how hard I flog it.

     

    Playback of DSD files with DSP and other added CPU stresses may push temp too high to ignore. If so, this would justify a case with good cooling (active or passive), but I haven’t yet run these scenarios and will provide my findings when I do.

    Share this comment


    Link to comment
    Share on other sites

    2 hours ago, bluesman said:

    The life of Pi is shortened by running often at temps close to the throttling threshold.

     

    Yes, the Richard Parker mod is not recommended.

    Share this comment


    Link to comment
    Share on other sites

    On 1/25/2020 at 10:29 AM, bluesman said:

    You’ll read more about temp control & overheating in the next chapter.  And there’s more on this in the following one (being prepared now), which is about hot rodding SBCs for audio with a focus on the RPi. Cooling is essential if you overclock, but a good fanless case like the Flirc (my favorite) is all you need for a stock Pi 4.  Placed in an area with good ventilation at an ambient temp below 25C (about 80F), even a stock plastic case should be fine for most audiophile use of a stock 4. I strongly recommend using heat sinks though - they’re tiny, cheap, and effective. The life of Pi is shortened by running often at temps close to the throttling threshold.
     

    The stock 4 throttles CPU speed down to that of your 3b+ when its temp hits 80C.  I’ve pushed one hard with live multitrack recording and real time monitoring, and it peaked at 73C in the standard plastic case. A Flirc brings that down to about 60C, and a $7 fan cooled case lowers it to a max of 43C no matter how hard I flog it.

     

    Playback of DSD files with DSP and other added CPU stresses may push temp too high to ignore. If so, this would justify a case with good cooling (active or passive), but I haven’t yet run these scenarios and will provide my findings when I do.

    Amazon, search:

    Jun-Electron for Raspberry Pi 4 Model B Case, Dual Cooling Fan, Enclosure Aluminum Alloy Case with Heatsink Pillar

     

    I've seen others similar, but passive only:

    Geekworm Raspberry Pi 4 Armor Case, Raspberry Pi 4 Computer Model B Armor Aluminum Alloy Passive Cooling Case Compatible with Raspberry Pi 4 Model B Only

    (Virtually the same case available from The Pi Hut for $15)

     

    A very detailed and credible article which is a good start to your series- looking forward to the next installment. I'm currently using J River for my audio and video needs, with a server and backup server, and put several small ITX FF computers I had lying around doing nothing into service as renderers in 4 rooms with a full version of JRiver on each, but sharing the server's library of roughly 800 blurays, 50 (and growing) 4k discs, 250 cds, and a growing collection of SACDs, all served at full original resolution, all on Win 7 SP1.

    I built my first computer in 1975 or 1976, am conversant in Windows, OS X, Linux and somewhat in Raspbian, which, under the hood have many similarities, which isn't surprising because they all have their roots in Unix. I would not dispute that some of your comments about OS X were true at some time, even within the last year or two, but newer tools are available and are continuing to be developed that make it easier than ever to build a Hackintosh that is enabled to do almost anything a Mac can do, including Continuity, Hand-off, Airplay, full Power Management, use of high-end GPUs, analog audio, bitstreamed audio over HDMI and DPI connection, USB, Firewire and Thunderbolt interfaces [selected compatible hardware required for all features]. It will cost you some time, research and patience. You can also opt for a build that doesn't do everything I just mentioned (many of which are not necessary for an acceptable audiophile-capable computer) and can be done in a couple of hours, depending on your previous computer skills. The first Hackintosh I built was with Snow Leopard (late 2009). The system was less complicated and took a lot more time. Today, the system is more complicated and takes less time. I build workstations because Apple hasn't built an acceptable workstation for my needs for quite some time, and the recently announced new Mac Pro is ridiculously over-priced, a bare-bones model costs $5.5k, a system minimally outfitted as a workstation costs over $15k, and maxed out it costs $53,658 US plus tax (all hardware is now proprietary form-factor, so you can't even buy GPUs from Nvidia or AMD to add on anymore). 

     

     

     

    Share this comment


    Link to comment
    Share on other sites

    On 1/25/2020 at 10:29 AM, bluesman said:

    You’ll read more about temp control & overheating in the next chapter.  And there’s more on this in the following one (being prepared now), which is about hot rodding SBCs for audio with a focus on the RPi. Cooling is essential if you overclock, but a good fanless case like the Flirc (my favorite) is all you need for a stock Pi 4.  Placed in an area with good ventilation at an ambient temp below 25C (about 80F), even a stock plastic case should be fine for most audiophile use of a stock 4. I strongly recommend using heat sinks though - they’re tiny, cheap, and effective. The life of Pi is shortened by running often at temps close to the throttling threshold.
     

    The stock 4 throttles CPU speed down to that of your 3b+ when its temp hits 80C.  I’ve pushed one hard with live multitrack recording and real time monitoring, and it peaked at 73C in the standard plastic case. A Flirc brings that down to about 60C, and a $7 fan cooled case lowers it to a max of 43C no matter how hard I flog it.

     

    Playback of DSD files with DSP and other added CPU stresses may push temp too high to ignore. If so, this would justify a case with good cooling (active or passive), but I haven’t yet run these scenarios and will provide my findings when I do.

    Not sure if you watched the video I linked. He said the Pi 4 sounds inferior to a 3 B+ for audio. Several comments below agreed with his assessment, which is the opposite of your recommendation.

    Share this comment


    Link to comment
    Share on other sites

    3 hours ago, audiobomber said:

    Not sure if you watched the video I linked. He said the Pi 4 sounds inferior to a 3 B+ for audio. Several comments below agreed with his assessment, which is the opposite of your recommendation.

    I've seen and read a lot of Hans B's output, and I simply don't agree with much (maybe even most) of what he hears and describes.  FWIW, his self-described background and qualifications to offer the opinions he holds are, to me, more than a bit specious:  "Before starting his carreer [sic] as a professional tech journalist, he sold hifi and recording equipment for ten years."  You'd think a "professional tech journalist" (even a self-described one) could spell correctly and/or proofread what he posts, which is not nit picking - if I paid him to write something for me, I'd expect it to be perfect on delivery and so (I hope) would you.  That directly quoted misspelling from his main website has wrankled me since the first time I saw it, and his other errors simply compound that.  If he's that unconcerned about detail and accuracy on his website (especially in his bio), why would we expect more accuracy in his audio reviewing? 

     

    I'm also more than a bit disappointed that he virtually never provides clear descriptions of what he hears or objective testing to support his contentions.  The presence of test equipment behind him in his pictures is misleading if he's not going to provide any results.  You're obviously of a different opinion, and that's fine - there's room for all of us, especially if we're civil in our expressions of disagreement.

     

    I'm not alone in my opinion about the Pi - in fact, I'm far from alone, and there is objective testing to support what I believe is the excellent SQ of a well set up Raspberry Pi 3 or 4.  For example, read this Pi 3b review with test data from Archimago.  Further, he's apparently unfamiliar with the USB bug in 4G Pi 4s right now, and it definitely affects audio.  I tested a few operating systems on a 4Gig Pi 4 to see if the USB problem is ubiquitous, and it's not.  Ropieee does not seem to display it - I've had no disconnects or sonic aberrations at all after 72 hours of connection.  Roon repeatedly drops and reconnects the Pi USB zone when running Bridge on Raspbian Buster.....but not Stretch, which works and sounds fine.  So there's an inherent problem in the 4 that is known to the development team, who are promising a patch "soon".  I only wish Hans had actually provided some technical info so we could know what he actually tested - was it Raspbian Buster, Stretch, or something else?  If I missed it, please let us know.

     

    Hans B offers 95+% unsubstantiated opinion, and I simply disagree with a lot of it (as do many others).

    Share this comment


    Link to comment
    Share on other sites

    9 hours ago, BluRay444 said:

    Amazon, search:

    Jun-Electron for Raspberry Pi 4 Model B Case, Dual Cooling Fan, Enclosure Aluminum Alloy Case with Heatsink Pillar

     

    I've seen others similar, but passive only:

    Geekworm Raspberry Pi 4 Armor Case, Raspberry Pi 4 Computer Model B Armor Aluminum Alloy Passive Cooling Case Compatible with Raspberry Pi 4 Model B Only.  Virtually the same case available from The Pi Hut for $15

    My favorite is the Kupton case with fan for $6.95 (with a 3A wall wart similar to the “official” Pi PS but with a switch in the line - it’s fine if you interrupt the power through USB and power your DAC directly & independently).  It comes with 4 little heat sinks and keeps an overclocked (2.3) 4 gig Pi 4 at an average of about 40C while recording live stereo or playing 2 ch DSD128 - the highest temp I’ve seen was a brief period at 43C.

    Share this comment


    Link to comment
    Share on other sites

    First, thanks to bluesman for an excellent read!
    However, I must admit that most of it went above my head. I found this article, because I am looking for the ideal setup to combine my Kef LS50W speakers with a new Samsung smart TV ("The Frame"). I currently have all my music on an external usb hard drive, and I have been using JRiver MC for some years and very much like that software. I thus read this article to find out if I should use a computer to connect these devices, and if so which one, or if I should buy some other device (the speakers are great at reproducing sound, but they do leave quite a bit to be desired regarding their streaming capability, so I was thinking that if some type of standard pc solution cannot solve my needs, than perhaps I need a proper server, which could be something like an Innuos mkIII?).

    I am still in as much doubt now as I was before reading this otherwise highly informative article, so If anyone here would be so kind as to help me with some advice as to how I can best solve my doubts and connect my devices for the optimal compromise between sound quality and convenience, I truly would be most grateful.

    My new tv has not arrived yet, and at the moment I am connecting my laptop directly to the speakers via usb, but I know it is far from the ideal solution.

    Share this comment


    Link to comment
    Share on other sites

    16 hours ago, bluesman said:

    I've seen and read a lot of Hans B's output, and I simply don't agree with much (maybe even most) of what he hears and describes.  FWIW, his self-described background and qualifications to offer the opinions he holds are, to me, more than a bit specious:  "Before starting his carreer [sic] as a professional tech journalist, he sold hifi and recording equipment for ten years."  You'd think a "professional tech journalist" (even a self-described one) could spell correctly and/or proofread what he posts, which is not nit picking - if I paid him to write something for me, I'd expect it to be perfect on delivery and so (I hope) would you.  That directly quoted misspelling from his main website has wrankled me since the first time I saw it, and his other errors simply compound that.  If he's that unconcerned about detail and accuracy on his website (especially in his bio), why would we expect more accuracy in his audio reviewing? 

     

    I'm also more than a bit disappointed that he virtually never provides clear descriptions of what he hears or objective testing to support his contentions.  The presence of test equipment behind him in his pictures is misleading if he's not going to provide any results.  You're obviously of a different opinion, and that's fine - there's room for all of us, especially if we're civil in our expressions of disagreement.

     

    I'm not alone in my opinion about the Pi - in fact, I'm far from alone, and there is objective testing to support what I believe is the excellent SQ of a well set up Raspberry Pi 3 or 4.  For example, read this Pi 3b review with test data from Archimago.  Further, he's apparently unfamiliar with the USB bug in 4G Pi 4s right now, and it definitely affects audio.  I tested a few operating systems on a 4Gig Pi 4 to see if the USB problem is ubiquitous, and it's not.  Ropieee does not seem to display it - I've had no disconnects or sonic aberrations at all after 72 hours of connection.  Roon repeatedly drops and reconnects the Pi USB zone when running Bridge on Raspbian Buster.....but not Stretch, which works and sounds fine.  So there's an inherent problem in the 4 that is known to the development team, who are promising a patch "soon".  I only wish Hans had actually provided some technical info so we could know what he actually tested - was it Raspbian Buster, Stretch, or something else?  If I missed it, please let us know.

     

    Hans B offers 95+% unsubstantiated opinion, and I simply disagree with a lot of it (as do many others).

    IME, technical qualifications have no bearing on someone's ability to judge sound quality. You do realize that English is Hans B's second language, right? I would have liked to hear your experience with SQ of the Pi 4 vs. Pi 3 rather than three paragraphs of Hans B flames.

     

    Greater computing power is not necessarily an advantage for music reproduction, and in fact can be the opposite. Measurements have limited value in assessing SQ. I expected the Pi 4 to be superior to a Pi 3 for playing music, but that is not what I've read thus far, except in your article. Below are a couple of expert opinions. 

     

    HiFiBerry says no advantage, but they do have a potential conflict of interest, since they sell Pi ancillaries.

    "If you’re using the Pi for music playback, there is no need to upgrade. We still know people using the first Raspberry Pi board that has been released 7 years ago. Music playback doesn’t require a lot of CPU power and/or bandwidth. Therefore, the music playback experience won’t be better with a newer Pi. Even older systems are powerful enough! If you still have a Pi1 or 2, use them. One advantage these systems have is the lower power consumption."

    https://www.hifiberry.com/blog/the-raspberry-pi-4/

     

    Dimdim is a computer engineer and DIY'er with no commercial interests:

    "We listened using the RPi3, then shut it down and booted up the RPi4, listening to the same material. Much to our surprise, we actually preferred the sound of the RPi3! The RPi4’s presentation had something of a “fatiguing” effect. The sound was a bit more “coarse” that that of the RPi3. We are not talking about big differences here, but they were there."

    http://www.dimdim.gr/2019/07/the-raspberry-pi-4-model-b-game-changer-for-audio/

     

    I am looking forward to reading your next article on the RPI, especially for suggestions on how to maximize SQ.

    Share this comment


    Link to comment
    Share on other sites

    4 hours ago, MeMa2010 said:

    First, thanks to bluesman for an excellent read!
    However, I must admit that most of it went above my head. I found this article, because I am looking for the ideal setup to combine my Kef LS50W speakers with a new Samsung smart TV ("The Frame"). I currently have all my music on an external usb hard drive, and I have been using JRiver MC for some years and very much like that software. I thus read this article to find out if I should use a computer to connect these devices, and if so which one, or if I should buy some other device (the speakers are great at reproducing sound, but they do leave quite a bit to be desired regarding their streaming capability, so I was thinking that if some type of standard pc solution cannot solve my needs, than perhaps I need a proper server, which could be something like an Innuos mkIII?).

    I am still in as much doubt now as I was before reading this otherwise highly informative article, so If anyone here would be so kind as to help me with some advice as to how I can best solve my doubts and connect my devices for the optimal compromise between sound quality and convenience, I truly would be most grateful.

    My new tv has not arrived yet, and at the moment I am connecting my laptop directly to the speakers via usb, but I know it is far from the ideal solution.

    And thank you - I really appreciate the kind words.  The "Value Proposition in Computer Audio" series is focused on pure audiophile use, so it looks in depth at critical issues for music listening (e.g. sound quality, music library management, album art display, tagging, external data search and retrieval), ease of use (e.g. downloading, installing, configuring, listening),  skill & effort levels required, and choosing a computer platform for your music software.  To this end, each article has a table or other summary suggesting the pros and cons of commonly used platforms for different music-related needs. 

     

    The introductory article was a general discussion of how computer based audio systems turn performance into playback. It's process based - for example, here's the front end article suggests some players and music management programs for different wants and needs:

     

    FOR

    CONSIDER

    FOR ITS

    Musicians

    Audacious, Sayonara

    Flexibility, effects like control over playback speed & pitch

    Jacket junkies

    Jajuk

    Comprehensive web sourced art & info displays

    Techies on budget

    Foobar2000

    Full features if you’re willing to work a bit to get them going

    Casual listeners who don’t want to have to fiddle with it

    GogglesMM, Gauyadeque, QMMP, Rhythmbox

    General layout, usability, stability, simplicity

    Web radio listeners & other internet streamers (from web servers, not yours)

    Cantata, Foobar2000, Kodi, Lollypop, Moode, QMMP, Sayonara

    Excellent setup, display and selection of web radio stations & streams;

    Streaming from your own server

    Foobar2000, MPD clients

    Ability to access your own music server remotely; Cloud services make this less necessary; takes some config work;

    Those with huge music libraries

    GmusicBrowser, MediaMonkey, QuodLibet

    Ability to digest, display, and manage lots of files

    Great tag management

    MusicBee, QuodLibet

    Tagging capabilities & flexibility

    A simple system you can make with little work and less $

    RuneAudio, Volumio on a Raspberry Pi

    Great value in simplicity, reliability, sound quality, etc;

    A simple player you can drop onto your PC

    Daphile, MusicBee

    Simplicity and ease of installation; versatility – play from computer’s audio out, drive USB DAC/amp etc; great & easy intro to computer audio

    Most of what you get from JRMC, Roon etc without the cost

    Foobar2000

    Amazing flexibility and continued development over many years; it takes some work to learn about plugins etc, and configuration isn’t always easy or intuitive – but you can get fantastic sound from almost any source with just a bit of work

     

    The series is organized to parse the various wants and needs into categories, to simplify the search for good solutions. Here's the first set of questions asked:

    • What computer will you use?

      • One you already have

      • A new one you plan to buy or build for multiple uses, one of which will be audio

      • A new one you plan to buy or build as a dedicated audio player

    • How complex do you want your system to be?

      • One box solution (player, server, etc)

      • Computer front end with NAS or other remote server

      • Other

    • What are your player access needs?

      • None – player to DAC to audio system at one location like the old days

      • Remote control only on LAN with mobile app

      • LAN – multiple endpoints / renderers on a home network

      • WAN – access to home server to stream remotely

    • What program sources will you access?

      • Local music files only

      • Web radio

      • Other internet streaming sources

      • Video material

    • What kinds of music files will you play?

      • mp3

      • High resolution FLACs

      • DSD

      • MQA

      • Multichannel

      • Other

    • What kind of computer will you use?

      • x86 etc

      • ARM based, SBC

      • Other

    • What operating system will you use?

    • Do you tend to stick with something you like for a long time or change around a lot?

    Your inquiry seems to address a comprehensive media hub, which is a different kettle of fish that we're not frying in this series :) But you can take the same approach I used to hone in on good answers to your questions.  The Frame looks like a cool device - but it comes with a lot of functionality that duplicates what you may already have, e.g. voice control, web and mobile control apps.  It has its own "assistant" (named Bixby) but can also be integrated with both Alexa and Google Assistant.  If you don't have a home hub already, the Samsung SmartThings Hub offers integration among your Samsung devices and can help you create a "smart home" if that's something you want. But you have to know what you want before you can figure out how to best get it.  So you have to define your wants and needs.  Here are some questions to get you started:

    • What do you want to be able to do when you're set up and running?
      • listen to recorded music
        • files on your own device(s) and/or networked storage
        • internet radio
        • streaming services
      • watch videos (which can also have different sources requiring different equipment & functionality)
      • have unified control over media selection and playback, "smart home" functions like lights, thermostats, etc
        • through a home hub
        • through an "assistant" like Alexa or Google Home
        • through a web and mobile app
      • other - there are many, many possible answers to this question & only you know what they are
    • What do you mean by "...[using] a computer to connect these devices"?
      • to which devices do you refer?
        • just the speakers and the TV
        • your current laptop, the speakers, and the TV
        • more or other devices
      • what needs connecting - data sources, media outputs, other?
      • do you have or want to set up a LAN, WLAN, or both?
        • do you want a networked media system that will let you play audio and video through the new TV?
    • Are you sticking with USB storage for your files or considering networked storage?
    • What do you mean by "combine my Kef LS50W speakers with a new Samsung smart TV"?
      • Do you want TV sound to come from your KEFs but be controlled (volume etc) by your TV?
      • Do you want to integrate your KEFs into a multichannel home theater setup?
      • Do you want to play multiple media from networked storage through your TV? HT?
      • Do you want to use JRMC as a true media center for your network or only for music?
    • What are your future plans for all this?
      • Are you building a system today for the long term with no plans to upgrade in the foreseeable future?
      • Are you building a platform on which to expand your network and/or media system(s) over time?
    • What are you doing about backing up your media files?
    • How big a chunk of your resources (time, money, space, structural mods like electrical lines / outlets / conditioning, etc) will you dedicate to this?
    • How much assistance will you need in designing your system, selecting the components, setting it all up, etc?

    This list could go on for days.  So you have to decide exactly what you want to have when you're done, which functions and factors are essential for your happiness, and what compromises you're wiling to make.  We have 4 Samsung "smart TVs" and each one has a different IQ.  I was amazed to learn (after I bought it) that our 2018 55" TV does not have the firmware to integrate with our Samsung SmartThings hub.  I'm no expert on smart homes and networked multimedia systems, but I've been a GE-certified Lean/6 Sigma master black belt for over a decade and I know a fair amount about process and system engineering.  Despite this, I failed to choose a new TV that was compatible with our desires.  If you don't define your needs up front, you'll never know if you met them.

     

    Good luck!!

    Share this comment


    Link to comment
    Share on other sites

    3 hours ago, audiobomber said:

    IME, technical qualifications have no bearing on someone's ability to judge sound quality. You do realize that English is Hans B's second language, right? I would have liked to hear your experience with SQ of the Pi 4 vs. Pi 3 rather than three paragraphs of Hans B flames.

     

    Greater computing power is not necessarily an advantage for music reproduction, and in fact can be the opposite. Measurements have limited value in assessing SQ. I expected the Pi 4 to be superior to a Pi 3 for playing music, but that is not what I've read thus far, except in your article. Below are a couple of expert opinions.

    Yes, I'm well aware of Hans' origins.  I'm not flaming him at all - I'm expressing the opinion that the value of what he offers seems far less to me than it is to you, and I'm explaining why.  It's very hard to believe that his command of English is less than excellent if he's done everything he put in his bio, e.g. having written "for a large number of magazines in the UK".  Further, it doesn't matter if English is his second language or if he doesn't speak it at all.  What matters is his level of attention to accuracy and detail (which, I hope you will agree, is critical for a tech writer).  He could hire a translator, an editor, a bilingual web designer, etc if he can't write up his material perfectly himself.  As a long time professional tech editor, he should be well aware of the rigors of publishing technical material and the many ways of achieving excellence in his field of expertise.  The errors in his content and presentation suggest to me that he's not as concerned about detail as he should be (or isn't as much of an authority as he claims to be).  I also generally disagree with what he says he hears. This is not denigration - many people disagree with Robert Parker's taste in wine and Roger Ebert's movie reviews, too.

     

    Hans was obviously either unaware of the current USB problem in Pi 4s or chose to ignore it.  He's in the same boat as the two experts you cite, neither of whom expresses awareness of the Pi 4 USB issue.  And if you read further down Dimdim's thread in which you found that quote, you'll find this from an otherwise anonymous contributor named Derek:

     

    "Well, it’s been a few days (and nights) of listening to the RPI4 installed in a PCB – previously occupied by a 3B – and powered by the same linear 1A supply as powered the 3B. Both running Moode. Both I2S. So, a swap of the 3B for the 4.
    Think I prefer the 4 in this configuration – linear PS and I2S out. Seems to have a hair more resolution, slightly better bass definition and a ‘feeling’ of more coherence. It’s not huge, it’s slight but I’m going to stick with the 4 in the unit."

     

    There are also many rave reviews of the Pi 4's SQ using USB, e.g. this one by David Snyder from September 2019. His bottom line for a Pi 4 driving a few different USB DACs (all with separate power supplies) is "Well, it’s early in my evaluation, but it sounds great so far. Like the Raspberry Pi 3 Model B+, the RPi4 has 5 GHz Wi-Fi and 802.11ac support, so streaming DSD256 and 32-bit, 768 kHz PCM are [sic] no problem. The background is inky black on quality recordings, leading to impressive dynamics."  I agree with this. Sadly, neither Mr Snyder nor the authors of many other such web posts include which OS is being used. Sadly, neither Hans B nor Snyder tells us if the Pi 4 evaluated is a 1G, 2G or 4G RAM model - and this matters because the USB issue is largely a problem of the 4G RAM variant.

     

    Higher CPU speed may actually affect SQ negatively for simple stereo playback, based on some experiments by Archimago and a few others. But CPU-intensive processes like DSP, transcoding, and sampling rate changes make a CPU work hard for a living. So pumping out DoP, DSD256 and other demanding playback modes can even affect SQ of some pretty hefty x86 devices when combined with DSP etc.  So you have to balance what you want to play and how with your choices of hardware and software to achieve the best compromise for your wants and needs.

     

    The next piece in this series is an overview and general discussion of small boxes / SBCs with only my summary impressions of the SQ of individual devices.  I'm currently working on the one to follow that, which is focused on the Pi from stock ZeroW, 3b+ and 4 to a hotrod 4 with overclocked CPU and GPU.  This will include more detail and links to files recorded, edited, and converted from format to format entirely on the Pi4. 

    Share this comment


    Link to comment
    Share on other sites

    Once again, a thousand thanks for your thorough reply, Bluesman! You just made me realize how much I have not thought about, but that I should think about.

    All I want is a decent, simple-to-use two-channel stereo setup centered around my Kefs. I hardly use any apps on my crappy and old (but still good enough for me) no-name mobile phone, and my main reason for getting a screen/TV is for managing JRiver. I never actually watch TV, and I haven't even had one since I moved out from my parents house some twenty years ago, nor do have I have any interest in a smart home (I would actually even prefer not having that, especially not having any microphones in the house, as that strikes me as being somewhat creepy). The reason for choosing "The Frame" is that it seems to be the aesthetically least intrusive screen that my wife and I have been able to find, since it can be put directly on the wall and almost looks like a work of art when not in use. We probably will get to watch a few movies on it too sometime on our external DVD drive or online, but probably not more than one or two a year, so its main purpose will be to enable the use of JRiver for music without having to open up the laptop. We did use to have a proper computer screen in our old house, but since we moved to another country a couple of years ago, we have been completely without screens apart from the laptops, and my old self-assembled i3 desktop computer has just been collecting dust. In our new and small home, a proper home office is however out of the question, and therefore so too is a traditional computer screen, as we agree that it would be ugly no matter where we would put it. Naturally, the idea now is that a screen will provide the image, while the sound shall be coming from the Kefs, so what we need is the connecting link to make it all possible.

    I do have a proper high-end headphone setup, but at least initially I do not intend to connect it to the Kef+screen-setup that we are planning for the living-room. I just need to find out how to get the most out these Kefs, and for this purpose I want to be able to connect my external usb drive with 3gb of WAV music files to some device which lets me see and manage these files, while preferably also being able to use the setup for different kinds of youtube videos and possibly some internet radio too, which is something that both my wife and I do enjoy from time to time.

    Sorry for all the text, but the market for these kinds of things seems to me an absolute jungle, and I feel completely lost in it, so I am glad to report that I have read your texts several times now, and feel that I am learning a lot from them, so once again thank you very much for that.

    Share this comment


    Link to comment
    Share on other sites

    5 hours ago, MeMa2010 said:

    Onmy main reason for getting a screen/TV is for managing JRiver. I never actually watch TV, and I haven't even had one since I moved out from my parents house some twenty years ago, nor do have I have any interest in a smart home (I would actually even prefer not having that, especially not having any microphones in the house, as that strikes me as being somewhat creepy). The reason for choosing "The Frame" is that it seems to be the aesthetically least intrusive screen that my wife and I have been able to find, since it can be put directly on the wall and almost looks like a work of art when not in use. We probably will get to watch a few movies on it too sometime on our external DVD drive or online, but probably not more than one or two a year, so its main purpose will be to enable the use of JRiver for music without having to open up the laptop.

    The Frame has a built-in media player. You could simply plug your USB drive into it and play music without JRMC, using your KEFs as the TV speakers. Or you could put JRMC on your i3 computer, use the Frame as your monitor via HDMI, and drive your KEFs with either the computer or the TV. You’d need a remote control if you don’t want to leave a wireless keyboard & mouse out.  Any inexpensive Android phone or tablet with a browser will work fine. I’ve bought 2 in the last year for under $50 USD on sale.

     

    Enjoy!,

    Share this comment


    Link to comment
    Share on other sites

    13 hours ago, MeMa2010 said:

    The reason for choosing "The Frame" is that it seems to be the aesthetically least intrusive screen that my wife and I have been able to find, since it can be put directly on the wall and almost looks like a work of art when not in use. We probably will get to watch a few movies on it too sometime on our external DVD drive or online, but probably not more than one or two a year, so its main purpose will be to enable the use of JRiver for music without having to open up the laptop.

    Another simple alternative is to get yourself a Raspberry Pi 4. You can use its HDMI port to drive the TV, connect your USB drive to it, and power your speakers with either USB or Bluetooth from the RPi or from your new TV. JRMC runs well on the 4, although you’d have to pay for that version separately even though you already bought the one on your laptop. An inexpensive tablet would work fine as a remote for that too - but you need a WLAN to use a mobile device as a remote in any system you set up. There is a way to set up a a Raspberry Pi as a WiFi hub, but I think you’re better off using a simple ac wireless router.

     

    JRMC on a Pi does not do quite as much as it does on an x86 PC. But for a simple 2 channel audio system, it’s quite fine.

    Share this comment


    Link to comment
    Share on other sites

    That is some very practical and useful advice. I now know what I can and want to do, so thank you very much for that. I am very grateful to you, both for your articles and your replies to me here. I will keep reading.

    Share this comment


    Link to comment
    Share on other sites

    Well, LOL! It took me a whole day to get my Pi4 to make music via usb to my W4S re-clocker and finally to my Bryston BDA-2 DAC. But, I am now happily playing Mahler's 2nd. It took me that long because I did all kinds of things wrong and did them wrong at least three times, then changed my mind about the OS a few times, and on and on. But I am very happy with the result. The only question I have is, is ALSA as good as WASPI or ASIO or what-have-you. But that discussion would be for another thread. Just know that I got it to work, and am looking forward to the next installment detailing SBCs. Thank you once again!

     

    John J

    Share this comment


    Link to comment
    Share on other sites

    15 hours ago, JJinPDX said:

    Well, LOL! It took me a whole day to get my Pi4 to make music via usb to my W4S re-clocker and finally to my Bryston BDA-2 DAC. But, I am now happily playing Mahler's 2nd. It took me that long because I did all kinds of things wrong and did them wrong at least three times, then changed my mind about the OS a few times, and on and on. But I am very happy with the result. The only question I have is, is ALSA as good as WASPI or ASIO or what-have-you. But that discussion would be for another thread. Just know that I got it to work, and am looking forward to the next installment detailing SBCs. Thank you once again!

     

    John J

    Congratulations and welcome to the club - you’re now an official Pi head!  Yes, ALSA is as good, with direct kernel-to-hardware communication and the drivers built into the kernel. But in the Linux sound platform, ALSA sits “below” Pulse and JACK, which are roughly the Linux equivalents of WASAPI and ASIO.  And, as in Windows, you don’t have to use them - but they’re there (if installed, as not all distros load both on installation) to help if you want / need them.

     

    Although you can use ALSA without Pulse, most distros install it if there’s any audio software at all in them. Pulse does not replace ALSA, it helps you manage it.

     

    JACK is a fancy patch panel with which you can “connect” the inputs and outputs of your Linux audio system, including MIDI and audio instruments and other sources, for complex recording and monitoring as well as mix down, mastering etc. I use it with Ardour for live recording, but you don’t need it just for listening to your music.

     

    There are other audio packages you’ll see but not need, like Carla. Linux audio can be complex if you let it or if you need that level of sophistication. But most audiophiles do not. Have fun!

     

    PS: what OS and music management system did you finally use?

    Share this comment


    Link to comment
    Share on other sites

    1 hour ago, bluesman said:

    PS: what OS and music management system did you finally use?

    PPS: Is your Pi the 1, 2 or 4 gig model?

    Share this comment


    Link to comment
    Share on other sites

     

    6 hours ago, bluesman said:

    Congratulations and welcome to the club - you’re now an official Pi head!

     

    PS: what OS and music management system did you finally use?

     

    I ended up with Raspian Lite downloaded from the raspberrypi.org website. I'm using Roon, so the Pi is setup with Roon Bridge. 

     

    5 hours ago, bluesman said:

    PPS: Is your Pi the 1, 2 or 4 gig model?

     

    RPi 4 Model B/4GB. Probably overkill just for Raspian Lite and the Roon Bridge, but I have it hooked up to the TV and will eventually load Kobi so I can play all my ripped DVD's that are stored on the NAS, along with my music. I must mention that where I have my system cannot be reached by ethernet or coax cabling, which irritates me no end. So everything is by wifi. I'm using the 5 GHz band and so far there have been no hiccups with this. The highest I can go with the Bryston is 24/192, but that plays with ease. I cannot do DSD, so have no experience over wifi with that. 

     

    Again, thanks to you for your great substantive essays and all the others who make this amazing hobby so accessible and fun. 

     

    John J

    Share this comment


    Link to comment
    Share on other sites

    1 hour ago, JJinPDX said:

    RPi 4 Model B/4GB. Probably overkill just for Raspian Lite and the Roon Bridge, but I have it hooked up to the TV and will eventually load Kobi so I can play all my ripped DVD's that are stored on the NAS, along with my music. I must mention that where I have my system cannot be reached by ethernet or coax cabling, which irritates me no end. So everything is by wifi. I'm using the 5 GHz band and so far there have been no hiccups with this. The highest I can go with the Bryston is 24/192, but that plays with ease. I cannot do DSD, so have no experience over wifi with that.

    I have a Pi4 in our living room running Bridge via 5GHz Wifi and it does DSD128 (actually DoP into a 2nd gen SMSL SU-8) just fine. If you're not already doing so, you might consider a good case for your Pi.  I have a Flirc passive case on the living room Pi and a fan cooled case on the hot rod with which I'm doing the work for the article on modding SBCs for audio (which will follow the next one, already completed & carrying more info on the Pi and other SBCs).

     

    If you were unaware of it, there's a bug in the 4 gig Pi 4b that limits RAM access - we're not able to use the full 4 gigs with the current firmware.  This model also has intermittent USB problems that go away if you limit memory access to 3 gigs in one of the configuration files.  The development team is working on a fix, and I'm hopeful that it'll be available soon.  Roon drops and reconnects one of my two 4s running Bridge on Raspbian Buster over a roughly regular cycle of about 30 to 60 seconds - but I don't have this problem with Ropieee.  I'm about to try with DietPi and Bridge to see if it's peculiar to Raspbian.

    Share this comment


    Link to comment
    Share on other sites

    19 hours ago, bluesman said:

    If you were unaware of it, there's a bug in the 4 gig Pi 4b that limits RAM access - we're not able to use the full 4 gigs with the current firmware.  This model also has intermittent USB problems that go away if you limit memory access to 3 gigs in one of the configuration files.  The development team is working on a fix, and I'm hopeful that it'll be available soon.  Roon drops and reconnects one of my two 4s running Bridge on Raspbian Buster over a roughly regular cycle of about 30 to 60 seconds - but I don't have this problem with Ropieee.  I'm about to try with DietPi and Bridge to see if it's peculiar to Raspbian.

     

    I did not know about the memory limitation. I did try to expand the use of RAM to its max, and apparently it did. Using "free -h" I see that I'm at 3.8Gi, but only using 202Mi with 3.5Gi free. The unit has been playing now for over 6 hours with absolutely no hiccups at all. Bridge is the only thing it's doing; I haven't loaded Kodi yet. I've also unplugged its power supply and plugged it back in several times to see if it will boot back up and be usable. No problems there. For my simple needs, it appears to be the solution. 

     

    I'm buying another and giving this one as a gift to an audiophile friend who has difficulty using a remote, so this will be perfect for him. The new one will have the flirc case. 

     

    I'm also documenting every single step I used to get this up and to its current state, in case someone wants to avoid all the trial and error and just wants to have a printout handy that is specific for the RPi4 and Bridge. 

    Share this comment


    Link to comment
    Share on other sites

    7 hours ago, JJinPDX said:

    I did not know about the memory limitation. I did try to expand the use of RAM to its max, and apparently it did. Using "free -h" I see that I'm at 3.8Gi, but only using 202Mi with 3.5Gi free. The unit has been playing now for over 6 hours with absolutely no hiccups at all.

    The USB and RAM problems with the 4GB Pi 4 are inconsistent - I have a USB problem on one of mine but not the others, and only with Raspbian Buster (even with the latest firmware). The problem is apparently in DMA (direct memory access) in 64 bit systems - here's a link to the technical description of the problem, and here's a link to the rapsberrypi.org forum thread about it.  I don't understand how you can access a full 4G of RAM with Raspbian Buster and maintain good USB connectivity, but this is a strange problem and I hope Canonical comes out with the fix soon.

     

    Here's a link to the official Ubuntu communication about it. In a nutshell, the Ubuntu team says:

    • "official support [from the Ubuntu team] for this board is currently limited to the 1GB and 2GB versions. Due to a kernel bug, USB ports are not supported out of the box in the official arm64 image on the 4GB RAM version. Kernel fixes have been identified by Canonical engineers. We are currently testing these fixes extensively. We will push updates within weeks, following successful test completion.  In the meantime, developers can use a temporary workaround to enable USB on the 4GB RAM version. The temporary solution will consist in editing the file /boot/firmware/usercfg.txt to limit RAM to 3GB"

    I'm using my problematic 4GB Pi 4 only as a test mule for now. Both the 3GB limit workaround on Raspbian and Ropieee work fine, so listening in our living room zone is uninterrupted until a proper fix is pushed by Canonical.

    Share this comment


    Link to comment
    Share on other sites

    I don't know if this is a problem or not, but I'm recording a temp of 68C using vcgencmd measure_temp. Seems high to me. 

     

    Still no hiccups though. Might try DietPi for grins and giggles. 

    Share this comment


    Link to comment
    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...