Linux audio is the way go, No its not

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
"So very much wrong!!"

MPD is a local headless "player" with several remote control options and an awful way to get it configured.

And that's about it.


A real and serious client-server system is a logitechmedia server and squeezelite. (all free)

And that approach is IMO ways better then MPD.

You can even easily integrate your tablets and phones (any OS) as playback "clients" by spending a little amount for the playback clients.

You can even sync those clients during playback.


This setup rocks. This is IMO by far the best network (client server) solution out there.

Cheers


PS:

Still. Linux comes with tons of flaws. It remains an OS for people who can accept basic functionality, buy HW that is supported by Linux and are willing
to work with the command line once in a while. Beside that Linux Audio is getting worse and worse with stuff like pulseaudio.


As much as I am a Linux guy myself, when it comes to applications I still prefer DBpoweramp/EAC, JRiver, MP3TAG on a Windows system.

As a special purpose animal (my "client-server" setup) I go the Linux route.
 
Linux is fine.

The packaging of desktop shells for distributions are getting worse and worse with things like pulseaudio. Who ever is putting together the packages seem to have forgotten the difference between recommends and depends.

It usually just tells us users that the politics upstream has overtaken technical ability and its time to find another desktop environment. For me the warning signs came when I couldn't disable the 'delete to trash'.

But this is a normal part of the Free ecology.
 
The packaging of desktop shells for distributions are getting worse and worse with things like pulseaudio. Who ever is putting together the packages seem to have forgotten the difference between recommends and depends.

A fair call. Canonical (for example) are turning Ubuntu into a "Linux for Windows Users" but aren't even getting that right. Pulseaudio is an evil thing but fortunately is easily removed. The problem is that for a newbie, it's easy to accept "that's how it is" and get frustrated and turned off. For the 'point and click' crowd, this ruins the Linux experience IMO.
 
pf..
windows vs linux.

linux is safer, harder to hack.
--> is that a priority within a mediacenter? nope..
back in the old days linux ment poor support and lack of drivers.
nowdays a lot of stuff is supported out of the box.
--> with windows drivers where never an issue. and will never be.
performance wise linux can outpreform windows.
--> does it mather? hard to find any computer that could not play video and/or audio files efforthless.

the main problem is, linux people thing users are intrested in the OS.
wrong. bloody not, users are not intrested in OS features and customisation.
Rather, they are intrested in the most troublesome installation of software and the wides compatibility possible, with the least time spendt with setting up the OS it self.
think about package dependency and other horrid stuff.

linux is good. when its the right tool for a job.
as a media center it can surely made to be verry efficient and what not.
but all this comes at a price.
a price most users are not intrested to pay.

point and click, something linux never got right.
surely nowdays linux managed to get a more r less decent GUI, but even then many times the good old console is needed. I personaly have no problems with it. but the majority does indeed have a quite good question, why the CLI, when it has a GUI ?
intuitive use. Well, nope. with linux one does NOT gona get intuitive use.
and this can be argued with. intuitive for those whom are used to it.
not for those who are not.
windows is intuitive in the majority of times even for those who are not used to it.

the software that plays the audio/video files can be identical on a linux machine and on a windows machine.
so the logical question is, why invest time into linux ?
I see no reason at all.
powercunsumption and such... oh my. intel atom platform or the amd eqvivalent is well suited to be a media center, they are low powerconsumption allready.

I'd only fo for linux if i was used to it, and would not mind the extra efforth to set it up properly and maintain it. other than that no way.
 
so the logical question is, why invest time into linux ?
I see no reason at all.

There are only one reason -> Fidelity are much better with a properly compiled linux kernel and properly compiled MPD player.

And then it works month after month or year after year.

I must admit that it took me a couple of months to patch, modify and compile a linux kernel and MPD player that I was satisfied with, but then I had to fight several idiotic bugs and modify the source code for both linux and MPD.


The years wasted on Windows with poor fidelity and much more time used just to keep in nearly working are now just an old memory :)
 
Last edited:
As much as I am a Linux guy myself, when it comes to applications I still prefer DBpoweramp/EAC, JRiver, MP3TAG on a Windows system.

Add CUEtools to that list.
I was very surprised that there almost was a comfortable alternative for linux called Rubyripper. One more of the promising projects that died slowly and silent.
A person called Grigory Chudov even showed interest to add Accurate Rip support into it when you read the bug mailing list but nothing happened anymore. He is now the person doing CUEtools.
Instead of any sign of progress on this one i lately heard about some other project named morituri that starts from scratch. What a waste.

btw. Windos 7 OEM licenses can be bought here in germany for 20-30€. I lately got another Pro license for 19.95€.
This is not even the money i cost for 1 our of my work. So needing more than one hour more fiddlingt to get things in a linux distro right against of Windows 7 is already a waste of money.
 
Administrator
Joined 2004
Paid Member
There are only one reason -> Fidelity are much better with a properly compiled linux kernel and properly compiled MPD player.
How can that be? I can get bit perfect playback and recording under Windows XP without much effort. I know, I've tested it. Then it's just up to the soundcard and its noise and jitter. That's a hardware issue, isn't it? Not software.
 
There are only one reason -> Fidelity are much better with a properly compiled linux kernel and properly compiled MPD player.

I am a Linux user and I dispute this statement. I have used Windows for audio and both operating systems can be tweaked to allow for high fidelity to the point of equality.

And then it works month after month or year after year.

Indeed it does. I think the last time my MPD server was rebooted was more than 6 months ago, and that was only due to a power outage.

I must admit that it took me a couple of months to patch, modify and compile a linux kernel and MPD player that I was satisfied with, but then I had to fight several idiotic bugs and modify the source code for both linux and MPD.

IMO if people stayed clear of the bleeding edge releases this shouldn't be necessary. Some apps are under constant modification and a bleeding edge release with a new kernel and libraries and old depends or code with missing features due to contstant updates or releases using the latest code from git are what may be causing newbies to Linux to pull the hair and return to Windows declaring "Linux sucks".
 
And at the end the majority of commercial HiFi Music Players out there are running Linux. Thought about that?!
Without Linux my diy headless HiFi Music Player Project would not be possible like that. And it only needs about 2-3 Watt. Controlled over Smartphone APP or WebUi or Buttons or IR remote.
 
How can that be? I can get bit perfect playback and recording under Windows XP without much effort. I know, I've tested it. Then it's just up to the soundcard and its noise and jitter. That's a hardware issue, isn't it? Not software.

It is the resulting transferred noise that affects the "sound card" and analog circuits that will give audible differences.

So YES the OS and drivers (software) matters as they creates noise in the PSUs. Due to that I have reduced and optimized the kernel, ALSA / USB drivers etc.. Also in the USB adapter there are optimized firmware and increased SRAM for buffering so the kernel - ALSA - USB drivers - USB adapter firmware and hardware can operate as optimal as possible with very low IRQ rates and optimal DMA transfers etc..

I am not talking about standard linux or MPD builds or a standard USB adapter.
Even the library that decodes a music file may have bad coding in linux and use twice as much CPU cycles as it should. I reported this as a bug, but this was not an issue that was going to be fixed - so I had to modify and compile an alternative library to optimize that also.

The MPD source code and config files was the hardest to modify to optimize as there was the greatest level of mismatch and hardcoding that was inconsistent.

Bitperfect playback are just one of the primary basic requirements.
 
I am a Linux user and I dispute this statement. I have used Windows for audio and both operating systems can be tweaked to allow for high fidelity to the point of equality.

Agree to this when standard Windows and standard Linux are used, but Linux can be modified and optimized by an extent Windows cannot.

IMO if people stayed clear of the bleeding edge releases this shouldn't be necessary. Some apps are under constant modification and a bleeding edge release with a new kernel and libraries and old depends or code with missing features due to contstant updates or releases using the latest code from git are what may be causing newbies to Linux to pull the hair and return to Windows declaring "Linux sucks".

The only issue with Linux and MPD have been the arrogance the developers have shown when I have been re-writing source code and config files, and have asked for help when there have been either clear bugs or inconsistencies that forced me to change much more of the code than I wanted.
Only WAV, AIFF and DSD are supported in my variants.
 
Administrator
Joined 2004
Paid Member
It is the resulting transferred noise that affects the "sound card" and analog circuits that will give audible differences.
I can see that, yes. At present I use an optical output so there isn't any noise moving from chassis to chassis that way. Not to say there isn't some already in the SPDIF signal.

As for stability, I haven't rebooted my Windoze music server in about a year. Of course, I use it for nothing but that purpose. I'm very surprised at how stable it is.

I'm hip to the lean, mean Linux music machine, and really didn't think Windoze XP would be the way to go. But so far it has just worked. Not much worry or bother.

One thing that does bug me in Windoze is power management. I have my computer set to sleep when not in use. In sleep mode it consumes very little power and takes just the push of a button to wake up. But it doesn't always go to sleep, and I can't figure out why.
 
Agree to this when standard Windows and standard Linux are used, but Linux can be modified and optimized by an extent Windows cannot.

I often read that and i don´t deny you have tons of switches you can push in linux but is there some place you can link me to where some serious testing of common audio hardware shows better performance in linux? I don´t mean some fuzzy feeling someone develops after pushing 1000s of knobs.
Besides that i´ll never use any computer to directly playback music for my main system listening, maybe feeding a DAC.
 
I often read that and i don´t deny you have tons of switches you can push in linux but is there some place you can link me to where some serious testing of common audio hardware shows better performance in linux? I don´t mean some fuzzy feeling someone develops after pushing 1000s of knobs.
Besides that i´ll never use any computer to directly playback music for my main system listening, maybe feeding a DAC.

I do not know of any testing that I could call serious.

And I have to add that I started with Linux on Intel platform - a triple boot system with Windows, OS X and Linux so I could compare with the identical hardware. In my opinion OS X and Linux was superior to Windows.
With a ARM based system and optimized Linux kernel, ALSA and MPD my opinion are that Linux are also superior to OS X.
When I get time I will finish a new ARM based music player setup where galvanic isolated I2S (and I2C for control) are taken directly from the ARM chip. The ARM I2S hardware are run in slave mode and galvanic isolated clocks are generated by the DAC master clocks and run synchronously. Data transfers in the ARM chip by DMA..
 
I'd be intrested to see some mesurements in audio quality of the same exact hardware and linux vs win xp vs win 7 .
I do not think the difference would be audible, but since i never tested it so i can not say i'm sure.
allso would be intrested in the same test with an external non computer DAC. so using only the digital output of the computer.
 
I'd be intrested to see some mesurements in audio quality of the same exact hardware and linux vs win xp vs win 7 .
I do not think the difference would be audible, but since i never tested it so i can not say i'm sure.
allso would be intrested in the same test with an external non computer DAC. so using only the digital output of the computer.

+1. that makes good sense!
 
I'd be intrested to see some mesurements in audio quality of the same exact hardware and linux vs win xp vs win 7 .
I do not think the difference would be audible, but since i never tested it so i can not say i'm sure.
allso would be intrested in the same test with an external non computer DAC. so using only the digital output of the computer.


Hi Arty,

I have done some of those measurements about a year ago, using Windows7 PC. The results are in http://www.bodziosoftware.com.au/Computers_SNR.pdf

Currently the best PC audio codes are ALC1150 with 115dB SNR.
You would be surprised how good PC audio is these days. With a proper selection of audio codec, you do not need any external hardware to have HiFi in your AV room.


Best Regards,
Bohdan
 
@Bohdan - I think you would run into problems with using those measurements to compare sound among audiophiles.

I did a session this weekend listening to windows 7 with hi-def playback through foobar2000 and Linux running Ubuntu and Clementine and Jack (non-real time) and found windows to sound better, then compared to ap-linux and running jack in real time and found ap-linux sounded better.

I don't think you can use measurements to determine what sounds better, as we don't know that what we can measure is what we are hearing. In this computer-usb-dac world though it does seem that jitter is an important measurement.

Using a USB-S/PDIF converter between my dac and my computer makes a huge difference, but introduces some tradeoffs. It is not easy. Still the sound coming out of my S/PDIF of my Marantz CD transport sounds better than the S/PDIF or USB coming out of my usb-spdif or my computer usb directly. So we have some learning to do here.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.