Linux

[Grez]

Do any of you guys use any flavour of the Linux OS?

If so, what players do you use and why?

I've never done any serious listening to music on my computer but I was just interested what people think and what they do.....

:)

[Vincent Kars]

I used Ubuntu for a while
In the end I dropped it
- no improvement in sound quality compared with Win
- not as technically challenging as I thought but not as easy to configure as Win
- players interface a bit down compared with Win, gives me the Win95 feeling

I tried a couple of players: http://thewelltemperedcomputer.com/Linux/Linux.htm

[dave2010]

I thought computer audio was quite simple until a month or so back, when I looked at a Linux mag (probably bought it out of interest, so could check). Since then I've discovered that the audio in other systems - such as Windows - is also fraught with problems. The problems are normally well hidden from end users.

I wouldn't expect there to be any great differences between a well designed player in Linux from any other system, though note the wording "well designed"! My experiences with Unix and Linux are that they run most programs much faster, and they have seemed much more capable for high speed communications, but whether that's really that important for music players I can't say. I'd have thought that it's not really going to make that much difference.

The issues I expect you'll want to consider will be user interface issues, including features such as whether track identification and cover notes are available. Many of the good open source projects are cross platform, though often developed on Unix or Linux systems. You'd probably need to talk to a developer to know whether there's any great advantage for the end user. Where you may get a benefit is that some of these developers are really innovative, and could be obsessive about getting good results, though it may depend on their approach. If they simply try to reconstruct existing commercial products, then they might produce software which is no better than existing software, and looks old - as noted above. On the other hand, they may manage to do something really new and innovative - though there could be rough edges. Personally I wouldn't mind rough edges in the interfaces if the sound quality is really good, but not everyone thinks that way, and it should be possible to have good sound quality and a good interface.

[Peter Galbavy]

Not really a direct reponse, but while I run Ubuntu on my netbook for convenience, my nerdness is focused on OpenBSD. For those who have seen mentions of the "BSDs" before but don't quite understand the relationship, the BSD family are NOT Linux derivatives or distributions. They have a much older pedigree and there are plenty of histories out there, so feel free to research.

There is currently a project underway in OpenBSD to revamp the audio subsystem, primarily by and for guys who want to play with their hardware properly and use OpenBSD for stuff like music production. There isn't really much of a website / wiki culture like there is in the Linux world but there is plenty of mailing list history out there. Look up stuff by Alexandre Ratchov.

OpenBSD is about security, code audit, (non-formal) correctness and standards compliance. There are plenty of areas where standards do not exist so there is scope to lead too. Performance and stability come out of this work so I see OpenBSD as an excellent base for a (once again non-formal) realtime system.

http://www.openbsd.org/ - release 4.7 is about to ship and is very small an install so have a go at some point :D

[HighFidelityGuy]

I've been tinkering with Linux on and off for the last couple of years and I now use Ubuntu on one of my PC's at work. Personally I find Windows easier to work with and more accessible but some distributions of Linux are quickly catching up in these areas, especially Ubuntu. Ubuntu can be installed and up and running on the internet etc very quickly and easily. It's also free of course so. You often don't have to worry about drivers etc, it just works. However, if you want to do anything more advanced than just browsing the web, sending emails and writing letters then it can get tricky. You can end up in a situation where you need to do stuff via a command line/terminal interface and you simply don't know the commands. So you have to start searching the web to find guides. This can get rather long winded and frustrating. However you encounter these problems much less often now than you did a couple of years ago.

I love the new "Ubuntu Software Centre" which is basically like "Add/Remove Programs" in Windows but you can choose from hundreds of free software to add. You just search for the type of software you want, pick a product and then click install. It automatically gets downloaded and installed. Very simple. This also ensures that you're using software that's guaranteed to be compatible. So you don't have to faff around downloading source code and compiling it your self.

I'd certainly recommend you try Ubuntu as it's free and probably the easiest route into using Linux. Good luck. :)

[Grez]

Thanks for the input guys.

I've now been a regular Linux user (Ubuntu flavour) for nearly two years since my old computer packed up and I bought this one which came with Vista installed. To cut a long story short, I had a nightmare with Vista and ended up needing to rescue the files from my old HDD with the live boot disk of Ubuntu.

Having been impressed with the live CD, I eventually installed Ubuntu and it's now my preferred OS. Although it can be a bit fiddly sometimes, day-to-day use was a much nicer experience than Windows... and it would drive my laser printer and my scanner which Vista wouldn't!

I currently use the Banshee music player, mostly, which has a sort of iTunes feel to it, but I've never used the computer for any serious music listening - just stuff bubbling away in the background while I get on with some work etc. I just wondered if there were any players that sounded better than any others.

Keep 'em coming.

...and thanks :)

[trailer]

Linux Mint is a flavour of Ubuntu and I have found to be better with finding drivers etc. It uses the Ubuntu software update repository as well as its own.

[swampy]

Been using Linux for about 6 years now, mostly Fedora. Having worked as a programmer and IT engineer the switch to Linux was a revelation for me at the time after using windows from the 3.1 days.

Win7 is very good, I run this for my synth stuff. Linux with Gnome GUI is my main OS and once you get past the initial learning curve of Linux it is very hard to go back to windows full time. I still find windows quite clunky, slower and too much hand holding compared to some Linux distro's.

[Krisbee]

I've used a lot of different flavours of Linux over the years and prefer KDE to Gnome. I've just installed the latest version of pclinuxos 2010 (32 bit) and also have sidux (64 bit) installed and a few others tested as virtual machines.

The media players you can use will depend whether you choose Gnome, KDE , Xfce, openbox or some other graphical desktop. But which sounds best will be as much to do with how you've configured ALSA and what back engines the player might use, and the audio source.

Play a CD in your PC and you might see the LED flash continuously or in bursts. For example, the amarok player in KDE uses Xine as its back engine and it reads a CD in burst by buffering data. In this case it is more like playing from a ripped file than other media players. If the source is streamed audio, or audio files on hard drive then this difference is less important.

You've not said if you are using a sound card or an external DAC for your PC based playback. Optimal configuration ALSA can make a noticeable difference to playback SQ. It is generally reckoned that anything that introduces sample rate conversion, mixing, or effects is going to be detrimental to SQ. Hence the preference to keep the number of software layers to a minimum and give exclusive use of audio to one application at a time if possible. This is getting harder to do if you use KDE for instance. With KDE4, the likes of pluseaudio and phonon get in the way. It's not so bad with other graphical desktops. You can always use the CLI to test if you can hear a difference in audio playback when using your audio hardware as directly as you can.

e.g.


mplayer -ao alsa:device=hw=0.0 mymusic.wav


In my case as I use an external DAC, it's recognised as device hw 0,0 by ALSA which is given by:


aplay -l
**** List of PLAYBACK Hardware Devices ****
card 0: default [USB Audio CODEC ], device 0: USB Audio [USB Audio]
Subdevices: 0/1
Subdevice #0: subdevice #0

If you are using a sound card then device hw may not be available and you'd have to use plughw. This is true of the envy24 family (m-audio 2496 and others) that needs bit depth padding.

[Vincent Kars]

A couple of month ago I came across the following post:


If anyone is using mpd on Linux with a 24-bit USB DAC (such as the Wavelength Proton), mpd recently incorporated S24_3LE support into the latest version in git. This means mpd can send audio straight to the DAC instead of processing in alsa first. This means no 32-bit conversion in mpd before being converted to S24_3LE in alsa. mpd converts straight to S24_3LE and outputs to the DAC if you specify hw:0,0. I'm noticing a definite improvement on my Audioengine 2A's. Once this change makes it's way to an mpd release instead of only being in the git repository, I bet audiophile mpd users take immediate notice of the change.

ggking7
http://www.computeraudiophile.com/content/New-mpd-feature-cleaner-signal

It must be possible now to play 24 using hw:

MPD has a good reputation, it is a server therefore can be run headless
http://www.thewelltemperedcomputer.com/Linux/Players/MPD/MPD.htm

[serendipitydawg]

I am a big fan of Linux Mint, but not really a fan of Linux.

It perplexes me.

Mint worked from the word go.Instant sound quality upgrade.

Same files same PC same DAC same cables same amp and speakers, even same volume settings.

Perplexed

[dave2010]

Mint worked from the word go.Instant sound quality upgrade.

Same files same PC same DAC same cables same amp and speakers, even same volume settings.

PerplexedInteresting this. How are you connecting the DAC? USB, Toslink or S/PDIF?

What sources are you using? Ripped CDs - 16bits/44kHz or higher resolution? MP3? MP3s might benefit from higher resolution soundcards, so if somewhere in your chain you've got an upgrade from 16 bit to 24 bit, that might partially explain the improvement. You shouldn't get an improvement on CDs if that's the case though, unless you have some HDCDs which effectively are 20 bit.

Other possibilities are that error handling and error correction is much better/faster, though that would only apply if you are actually getting significant errors on your links. If, for example, you are using USB, then you might get the same improvement by using a much better USB cable, which should reduce any errors to close to zero. Some will advise you that Toslink could also be subject to errors, though personally I'd doubt that there'll be many errors if you are using Toslink with reasonable cable, and your cable isn't too long. Some people prefer coax for S/PDIF, but again I wouldn't expect too many errors unless you use really poor cable.

If you are getting errors on the links, it is possible that the Linux OS is handling the errors much more rapidly, and hence better, and this could be the reason for the noted improvement.

Just possibly the audio handling could be doing something even more radical, such as upsampling (or not) or passing the audio data through some digital processing, such as an interpolation filter. If your files are ripped from CD (DVD?) etc., did you use a tool such as EAC, or could there be errors in the files? If so, there could be some nifty processing in the Linux system which might minimise the effects.

For MP3 there are decoders, such as the MAD decoder which claim to provide better quality sound. I personally think there is something in that, as it does some improved dithering, and also I thought on several tests that it did sound more musical. Is it possible that your Linux system uses a better decoder for the sound sources you are using? An odd one I read very recently was a system which decoded FLAC files better than WAV, even though the files can be shown to have the same data.

Can't think of anything more right now. If these don't nail the problem down, then for now I, like you, will be perplexed.

[serendipitydawg]

Thanks for responding Dave2010.

Connection between PC and DAC is USB to an m-Audio Transit then Toslink (only choice available) to my Beresford TC7510.

Sources are ripped CD's using EAC, accuracy either 99.9% or 100%. In my view MP3's have nothing to with hi-fi; a degraded product if you will.

I forgot to mention, I am a fully paid-up sceptic. So I would initially be very sceptical that a USB cable could make any diffference to sound quality. The sceptic's view must surely be that the data either gets through or it doesn't.

I am equally sceptical that my rips are full of errors that don't sound as bad when played via Linux Mint. All done with EAC and show accuracy of 99.9% or 100%.


I have perfectly adequate SP/DIF cables coax and Toslink; both are as supplied from the mighty Mr Beresford.

I am still perplexed.

[Krisbee]

I was under the impression that using the m-audio transit in Linux was a bit fiddly as you have to load a firmware file to get it working. Perhaps newer Linux distros have got round this problem or have at least automated the process.

But when you say “Mint worked from the word go.Instant sound quality upgrade.” what are you comparing Mint with? Were you previously using another flavour of Linux, or Windows? As you ripped CDs with EAC it could have been Windows. You didn't say if you ripped CDs as wav, FLAC or something else. If you were using windows what media player did you use and had you set up ASIO or WASAPI? And what media player are using now (rhythmbox?) and have you taken an special steps in configuring ALSA or just have you used Mint's default settings?

To my ears audio play back in Windows without ASIO/WASAPI is inferior in quality to a standard/typical ASLA config in Linux.

[dave2010]

I forgot to mention, I am a fully paid-up sceptic. So I would initially be very sceptical that a USB cable could make any diffference to sound quality. The sceptic's view must surely be that the data either gets through or it doesn't.

I am equally sceptical that my rips are full of errors that don't sound as bad when played via Linux Mint. All done with EAC and show accuracy of 99.9% or 100%.Great. I'm glad that there's another sceptic here. However, sometimes this can go too far. For years I didn't believe that upgrades to cable made much difference. As I recall I tried new loudspeaker cable (not much difference there), and then some different analogue interconnects (this was many years ago). One recording just came to life with an upgraded interconnect - I was really surprised. I'll be shot down here, but I'm not convinced that there are vast differences in interconnect cables, but there's a lot of difference between the grotty ones which are often supplied with kit and ones a bit higher up the scale.

Regarding your current issue, I found several pages extolling the virtues of Kimber USB cable, but at a ridiculous price. I also found at least one guy suggesting that it was vastly overrated - here it is - http://www.head-fi.org/forums/f46/high-quality-usb-cable-usb-dac-353613/ I'm sure that there can be differences between USB cables and a really cheap one could give lots of errors. However you may not have to pay much more (probably less than a tenner) to get a much better one with screening etc., which would give good results.

I had your view regarding data on USB cables, but actually audio data is a bit different in the way it's sent between most devices. If large amounts of data are/can be buffered in the receiving device, then protocols such as TCP/IP can be used, and then the "bits is bits is bits" argument holds, and the quality is dependent on the receiving device. Most audio devices are not connected that way, and use different USB protocols, in which case error detection, correction etc. may all suffer. Either there'll be no error detection and correction, in which case it's important that all the bits get through without any errors, or if there is any error handling it may be relatively primitive, and delays in communication may also be a factor. The point is that audio does not normally use bulk mode, but other real time modes.

I'd guess that something like this cable - http://www.amazon.co.uk/Belkin-PRO-Hi-Speed-Device-Cable/dp/B000EOTDPU/ref=sr_1_3?ie=UTF8&s=electronics&qid=1271403810&sr=1-3 - would do the trick (less than five quid), but some will try to persuade you to buy the Kimber or other esoteric cables. I'm largely with you on the scepticism though.

Any more ideas about why you get the sound improvement? You're not imagining it are you? You still believe yourself?

Re MP3 and similar - again I'm in agreement with you that it's not great - but sometimes it can sound surprisingly tolerable - though it depends on the musical material and the encoding, and also the decoder. I use Spotify and Napster quite a lot because accessing different music is of interest to me. Sometimes the distortion is horrible, but sometimes things really don't sound too bad, and I get to hear the music.

So you rip CDs and you use EAC - so most of your data is accurate 16 bit.

I did think of one other possibility - clutching at straws here - but maybe the number representations in the Linux system are different. If the player uses floating point the data sent out might be different from a system which uses integer arithmetic. Of course the data should simply be transmitted without change, but if there are any digital filters in the player then the arithmetic capabilities could make a difference.

Another possibility is going back to jitter - maybe the calculations in Linux are so much faster that jitter on the link is reduced. I'd reckon that Unix/Linux systems are generally so much faster (maybe also better) when it comes to communications, that that could have something to do with the observed change.

Otherwise I might still be very perplexed!

[swampy]

Thanks for responding Dave2010.

I am equally sceptical that my rips are full of errors that don't sound as bad when played via Linux Mint. All done with EAC and show accuracy of 99.9% or 100%.

I am still perplexed.

Linux has an app called cdparanoia. It is command line but there are a few GUI's that use it such as Grip. It is as good as EAC but you can run EAC via wine I guess.

I also find windows sound worse via a usb DAC when not using ASIO drivers. Hence my early post in this section.

[HighFidelityGuy]

I'm also interested to know what operating system people moved from when they heard an improvement with Mint Linux. If the move was from XP then I can understand this as it's notorious for not producing a bit perfect output. Vista and Win7 on the other hand should use WASAPI to get bit perfect, so that should be as good as it can get I think.
Unless you didn't have WASAPI enabled or something. :scratch:

[Vincent Kars]

Comparing various OS is a bit hard to do IMHO.
If you have them on different boxes you introduces all the differences in hardware too.
If you have them on the same box, you must reboot to switch so you have a large time lapse in your AB.
If you use Virtualisation software you probably can do a rapid AB but what is the impact of adding this layer?
For each OS you must know how to configure it correctly, one parameter wrong makes your comparison useless.
A famous one is playing Redbook audio in iTunes on Win. By default it upsamples to 96. If you have Win7 set to 44.1, it will downsample the 96 to 44.1. You sure will hear a difference between iTunes and WMP in this scenario!

[Krisbee]

Well, it's not so hard if you happen to have two Pcs built with the same spec both with coax digital out and one of Stan's DACs which gives you easy switching between the two sources. Virtualisation software would be useless, run a Windows VM on Linux and all audio will still pass through ALSA etc. Otherwise I agree that the time lag between rebooting on a PC running multiple OS is a hindrance, but it depends on how well you know your test piece and if you feel confident you can distinguish differences in SQ. The example you give is of a difference between media players on the same OS, not between media players on different OS, so should be easy to detect. Certainly it is up to the user who wishes to obtain the best audio SQ from PC play back to educate themselves about the optimum configuration of the software they use. But I still contend that a standard/typical Linux config, which can be tried by using any number of LiveCDs , will be better than Windows without ASIO/WASAPI.

For the truly obsessive in Linux you can always kill your graphical environment and drop back to the Linux text console and play audio from there with a minimum of other services and processes running. This is not so easy to achieve in Windows, but in Linux it takes just one command,
init 3. Long live the Penguin.

[Vincent Kars]

The purist probably don’t want to have any screen at all (RFI).
Running a headless Linux box using MPD is one of the more popular solutions.

[trailer]

The purist probably don’t want to have any screen at all (RFI).
Running a headless Linux box using MPD is one of the more popular solutions.

Is there an app for that? :eyebrows:

[Vincent Kars]

Sure: http://mpd.wikia.com/wiki/Music_Player_Daemon_Wiki

[serendipitydawg]

I was under the impression that using the m-audio transit in Linux was a bit fiddly as you have to load a firmware file to get it working. Perhaps newer Linux distros have got round this problem or have at least automated the process.

But when you say “Mint worked from the word go.Instant sound quality upgrade.” what are you comparing Mint with? Were you previously using another flavour of Linux, or Windows? As you ripped CDs with EAC it could have been Windows. You didn't say if you ripped CDs as wav, FLAC or something else. If you were using windows what media player did you use and had you set up ASIO or WASAPI? And what media player are using now (rhythmbox?) and have you taken an special steps in configuring ALSA or just have you used Mint's default settings?

To my ears audio play back in Windows without ASIO/WASAPI is inferior in quality to a standard/typical ASLA config in Linux.

Chris: I was very suprised that my Transit "just worked". Mint has a dropdown menu for audio settings and there it was "M-audio Transit (OSS)"


Thanks for the major clue "without ASIO". I realised that I had long since given up on making the ASIO driver supplied with my Transit work properly.


My original post was prompted by the improvement I heard between non-ASIO Windows XP SP3 playing FLAC with foobar2000; & a standard Mint Linux install using the included Rhythmbox player playing FLAC.

Now I have ASIO functioning and I am less perplexed.

Dave: I relaxed my scepticism a little and tried a different USB cable. Inconclusive so far, but definitely worthy of further study.

[dave2010]

I relaxed my scepticism a little and tried a different USB cable. Inconclusive so far, but definitely worthy of further study.I am also testing some different USB cables, but it could well be sod's law and I may eventually come to the conclusion that the one I started with was the best one. No way I'm spending £40 on a USB cable.

[Vincent Kars]

No way I'm spending £40 on a USB cable.

40?
Locus Design Cynosure USB cable, starting at $2849 for 3FT. :stalks:
High End has discovered USB audio.

[serendipitydawg]

"No way I'm spending £40 on a USB cable. "


Is that how much a Kimber USB cable costs?

Will anyone sell me one with a money back guarantee?

[dave2010]

"No way I'm spending £40 on a USB cable. "


Is that how much a Kimber USB cable costs?

Will anyone sell me one with a money back guarantee?It seems you can get a 1/2 metre cable for around £35 - http://www.russandrews.com/product-NEW-Kimber-USB-cable-2800.htm The price goes up to around £50 for 2 metres, and £60-ish for 3 metres. There may be other sources.

I got a couple of cables recently - one 3 metres, and one 1.8 metres - total cost under a tenner. They seem to work OK.

[Vincent Kars]

Anybody tried this one: http://www.wireworldcable.co.uk/categories/usb_cables.html

[nbaptista]

40?
Locus Design Cynosure USB cable, starting at $2849 for 3FT. :stalks:
High End has discovered USB audio.

he,he,he :guitar:
BTW,I have a Kimber USB cable.

[dave2010]

he,he,he :guitar:
BTW,I have a Kimber USB cable.I'm sure these are very good, and do what they say they do, just like the tree in my father-in-law's garden.

He had a rowan tree, which he said was there to keep witches away. It worked. We never saw any witches.

I think a lot of high priced "hi-fi" stuff might be like that. They take's your money, and they keep the witches at bay - but they've got your money!

Cognitive dissonance - see http://en.wikipedia.org/wiki/Cognitive_dissonance - may have a role to play here, though note that it can be played in more than one way. Some, such as myself, justify my belief that cheaper cables are good enough by asserting that I can't hear any difference between my cables and others, while others, who buy expensive cables will justify their purchases by insisting that there is a difference, or that there is some other difference, such as the physical quality of the cable (which may have no bearing on the sound), or the aethetic appearance (e.g colour), or the gold plating on the plugs and sockets (which may have little real impact).

My view is still that there could be a difference between very cheap cables and some which are less cheap. However I doubt that it will really be possible to tell the difference in a short USB cable between one which is acceptably good and one which in relative terms costs a fortune. If you're rich, and don't care then go for the very expensive ones, but otherwise save your money.

[Alex_UK]

Great post Dave - and one I agree very much with - I've never compared super expensive (or even moderately expensive) USB cables, so I have no personal knowledge of them, but I suspect there would be little or no difference between my "acceptably good" Belkin Gold USB cable and a "super-cable" - if nothing else, I doubt my system would be good enough to reveal the differences (if they exist...)

I haven't got any witches either, since the 1st divorce! :)