soundcheck said:
I've started a wiki.
Have tried to start it off with the basics, as many here may be put off by moving. The simpler it is, the more people will end up with Linux set-ups -- which is good news.
But it's a wiki, so feel free to add, chop and change.
http://www.diyaudio.com/wiki/index.php?page=LINUX+Audio
ssmith said:
ssmith.
Great start!! ;-)
I'll fill in some more information over the weekend. I need to get them structured first.
Do you know if it is possible to store certain files over there?
To keep it simple for newbies it might be a good idea to make pre-compiled binaries or ready made scripts available, which of course bears a security risk!
I also think it would be a good idea to open up a seperate Linux-Audio-Wiki thread to cover imprvment proposaly and other Q&As.
BTW:
I just got feedback of my "first officially pimped" linux-audio-buddy. At this stage he's just running a tuned "XMMS" only.
To summarize it:
Lots of things he never heard before, much better low end, much better transients, asf.
So far so good. I didn't expect anything different, since it has been proven on a 2nd reference system before!
I hope that within the next couple of weeks we'll have some more systems up'n running!
Cheers
Klaus
Hi Klaus -- no problem -- being a linux newcomer I have nothing to contribute to the tech side of things, so am happy to at least try to help put something together that'll make it clear for others (and me!).
I don't think it is possible. What kind of file size are you talking about though? Perhaps I can put them on my webspace. We could link to the files from the wiki, and include a super-simple how-to explanation.
That would be a great idea once the wiki gets going.
I have some time tonight and tomorrow to get ubuntu-studio installed and up and running, so will hopefully be with you shortly and ready to experiment....
cheers, stefan
Hi there.
I fixed the first draft of the Wiki.
http://www.diyaudio.com/wiki/index.php?page=LINUX+Audio
You won't believe it, this exercise took me while!
Perhaps somebody can have a look at it.
There are some small outstanding issues of course.
brutus, my brutefir based player, I'll most probably introduce
next week.
Now we need a volunteer to run the exercise of checking if everything works as
described!
I am off for the rest of the day.
Cheers
\Klaus
I fixed the first draft of the Wiki.
http://www.diyaudio.com/wiki/index.php?page=LINUX+Audio
You won't believe it, this exercise took me while!
Perhaps somebody can have a look at it.
There are some small outstanding issues of course.
brutus, my brutefir based player, I'll most probably introduce
next week.
Now we need a volunteer to run the exercise of checking if everything works as
described!
I am off for the rest of the day.
Cheers
\Klaus
Hi Klaus,
Great work. Will give it a go as soon as I get the chance.
I managed to install ubuntu studio. Last time I gave Linux a try was back in 2000, which was Mandrake 8.0 or something. A few years ago I tried Mepis... But I can see it has come a long long way. It's been a real pleasant surprise -- very very easy, of course following the step-by-step install guide I linked to above.
The biggest surprise, however, is the sound. Frankly I'm speechless. I wasn't really expecting a difference, because all the Windows tweaks I'd done before changed things a little, sometimes a bit better, sometimes a bit worse.
Ubuntu-studio is a very big improvement in sound, and that's just using FLAC through whatever music player I managed to get working first (it happened to be rhythmbox).
cheers!
Great work. Will give it a go as soon as I get the chance.
I managed to install ubuntu studio. Last time I gave Linux a try was back in 2000, which was Mandrake 8.0 or something. A few years ago I tried Mepis... But I can see it has come a long long way. It's been a real pleasant surprise -- very very easy, of course following the step-by-step install guide I linked to above.
The biggest surprise, however, is the sound. Frankly I'm speechless. I wasn't really expecting a difference, because all the Windows tweaks I'd done before changed things a little, sometimes a bit better, sometimes a bit worse.
Ubuntu-studio is a very big improvement in sound, and that's just using FLAC through whatever music player I managed to get working first (it happened to be rhythmbox).
cheers!
Hi Stefan.
Below tweaks you can start with, even without having a new kernel compiled.
Configure the RAMDISK and XMMS recommendations without compiling the realtime kernel. Strongly recommended!
Very important if you have an Usb-DAC is the ALSA setting in chapter 4.
I any case you should apply the PCI latency tweak if you run a PCI bus.
BTW:
How do I access the Wiki from DIY-Audio Wiki Home? The Wiki area doesn't really look structured at all!
Cheers
Klaus
Below tweaks you can start with, even without having a new kernel compiled.
Configure the RAMDISK and XMMS recommendations without compiling the realtime kernel. Strongly recommended!
Very important if you have an Usb-DAC is the ALSA setting in chapter 4.
I any case you should apply the PCI latency tweak if you run a PCI bus.
BTW:
How do I access the Wiki from DIY-Audio Wiki Home? The Wiki area doesn't really look structured at all!
Cheers
Klaus
are you using the GNOME desktop or a light weight WM
I setup Ubuntu Studio and use XMMS in conjunction with the Blackbox WM. Do you think that bypassing the Gnome desktop and Gnome audio mixer, sticking with the alsamixer and using lightweight WM adds to the overall efficiency of the linux based transport?
I have not patched the kernel for realtime according to your wiki, but out of the box it sounds fantastic.
I setup Ubuntu Studio and use XMMS in conjunction with the Blackbox WM. Do you think that bypassing the Gnome desktop and Gnome audio mixer, sticking with the alsamixer and using lightweight WM adds to the overall efficiency of the linux based transport?
I have not patched the kernel for realtime according to your wiki, but out of the box it sounds fantastic.
Re: are you using the GNOME desktop or a light weight WM
Hi there.
I am actually not using any windows manager.
I boot up in recovery mode and run my own ASCII based player.
Doesn't matter what windows manager you use, it'll steal quite
some processing power or better "interrupt priority" from your audio application.
However.
Blackbox or XFCE are in any case preferable over Gnome or KDE.
Alsamixer as well as Gnome mixer are using ALSA dmix.
For best sound you have to get dmix out of the loop.
XMMS supplies AFAIK an internal software based volume control.
If you want to use a SW volume control, use this one and access your HW directly ( choose HW0 or HW1 as direct accessible device).
( This is similar to ASIO under Windows) If your volume control
under Gnome doesn't work any longer, you know that dmix is out of the loop.
Cheers
stevevarga said:
Hi there.
I am actually not using any windows manager.
I boot up in recovery mode and run my own ASCII based player.
Doesn't matter what windows manager you use, it'll steal quite
some processing power or better "interrupt priority" from your audio application.
However.
Blackbox or XFCE are in any case preferable over Gnome or KDE.
Alsamixer as well as Gnome mixer are using ALSA dmix.
For best sound you have to get dmix out of the loop.
XMMS supplies AFAIK an internal software based volume control.
If you want to use a SW volume control, use this one and access your HW directly ( choose HW0 or HW1 as direct accessible device).
( This is similar to ASIO under Windows) If your volume control
under Gnome doesn't work any longer, you know that dmix is out of the loop.
Cheers
Hi Klaus,
Managed to follow your how-to without problem, really appreciate your efforts, thanks.
Im running ubuntu server without a graphical interface, timer set to 5000hz (will try 10 soon) and using music player daemon with the schedtool settings of the how-to. Listening on a griffin imic I believe there is an improvement in detail from running stock xubuntu. The system boots in a flash even on a 700mhz cyrix (I will not be compiling again on this machine - took all day!).
MPD has some interesting features - the audio buffer can be set up to be large enough to hold a whole wav file and completely fill before a file is played - I am assuming this is comparable to setting up a ramdisk as the hard drive is not accessed during play.
I have been using ncmpc which is a command line interface that has all the browsing and playlist capability of a graphical interface, because mpd runs as a system service you can quit the interface and audio continues playing. I also tried contolling through the graphical client sonata, using fluxbox, and I have to say at moment I can not detect any change in audio quality with the addition of the memory overheads of the graphics. However, I'll compare again when I have something better than an imic.
Would be interested to know how to set up brutefir as a volume control.
I'll report back once I have done some more listening comparisons and put together a proper usb dac.
Cheers,
Ross
Managed to follow your how-to without problem, really appreciate your efforts, thanks.
Im running ubuntu server without a graphical interface, timer set to 5000hz (will try 10 soon) and using music player daemon with the schedtool settings of the how-to. Listening on a griffin imic I believe there is an improvement in detail from running stock xubuntu. The system boots in a flash even on a 700mhz cyrix (I will not be compiling again on this machine - took all day!).
MPD has some interesting features - the audio buffer can be set up to be large enough to hold a whole wav file and completely fill before a file is played - I am assuming this is comparable to setting up a ramdisk as the hard drive is not accessed during play.
I have been using ncmpc which is a command line interface that has all the browsing and playlist capability of a graphical interface, because mpd runs as a system service you can quit the interface and audio continues playing. I also tried contolling through the graphical client sonata, using fluxbox, and I have to say at moment I can not detect any change in audio quality with the addition of the memory overheads of the graphics. However, I'll compare again when I have something better than an imic.
Would be interested to know how to set up brutefir as a volume control.
I'll report back once I have done some more listening comparisons and put together a proper usb dac.
Cheers,
Ross
A soundcard whose sound quality changes depending on irrelevant parameters (like system load) is not "better" ; in fact it has worse noise/jitter immunity than a soundcard that doesn't show this behaviour.
Same for a DAC that allows to hear differences between transports : it only shows poor jitter rejection, not that "transparency" crap.
rossco_50 said:
Hi Ross.
Good to see you were able to follow the Wiki.
Especially on machines like yours it is worthwhile to also tweak the IRQ
priorities. I'll add this tweak later to the Wiki.
1. Did you also try XMMS as I recommended? XMMS supports internally
realtime operation. (I'll have a look at ncmpc - didn't know it)
2. Did you make sure that you have direct access to the soundcard/DAC.
You need to avoid dmix in any case?
3. My kernel compiles now in 9 minutes ( I started with approx. 50minutes) .
I've thrown out almost 90% of all modules.
4.
peufeu is of course right, with his statement. The PCM2707/6 is not
at all immune against jitter, although it's one of the most used USB
receiver chips in the usb-audio world I guess.
However. A question pops up about - What tells you that your soundcard is not that bad that you can not even realize the jitter induced on the way from the file to the soundcard?
As a matter of fact, even the best reclocker DACs respond to improvements further upstream.
(I wonder if peufeu ever tried at least some of the tweaks, since he is also using linux! I'd prefer if he would tell us that all of these tweaks do not have any impact on his ethernet setup in particular, which in fact might proove?!? the superiority of his solution
BTW: The 3rd system ( also with USB DAC) is now 100% running incl. brutus.
What I heard so far - makes me believe that I am on the right track with
my approach.
One more:
Next weekend I'll face the first rather official comparison of my Linux
solution versus the best Microsoft solutions on IMO one of the best audio-systems
money can buy. I'll report back.
Cheers
\Klaus
Klaus,
1. I installed mpd first as I didn't have a GUI - I do now so I'll try xmms. The realtime option of xmms was something I meant to ask you about. I tried to set up a ramdisk just now following your instructions, but I get the message "mount: /tmp not mounted already; or bad option" any ideas? I reduced the size to 100mb as I only have 512mb, so there should be enough space.
It would be great if brutus could be made to work with the mpd frontends - ncmpc provides the functionality without bloat.
2. MPD is set up to output directly to HW:1, the mixer doesn't work.
3. Next time I will spend more time looking at the modules I need.
4. I would say my last post, if anything, I was in agreement with Peufeu - the graphical load did appear irrelevant to my systems performance. However, I do consider the dddac to be much better than an imic, therefore until I have tried a dac comparable to the dddac I do not dispute Klaus' findings.
I have another motive to use the commandline on a older machine - even xfce made the system sluggish at points, with the command line only it is much faster. Fluxbox is also an improvement over xfce.
Look forward to hearing the competition results.
Ross
1. I installed mpd first as I didn't have a GUI - I do now so I'll try xmms. The realtime option of xmms was something I meant to ask you about. I tried to set up a ramdisk just now following your instructions, but I get the message "mount: /tmp not mounted already; or bad option" any ideas? I reduced the size to 100mb as I only have 512mb, so there should be enough space.
It would be great if brutus could be made to work with the mpd frontends - ncmpc provides the functionality without bloat.
2. MPD is set up to output directly to HW:1, the mixer doesn't work.
3. Next time I will spend more time looking at the modules I need.
4. I would say my last post, if anything, I was in agreement with Peufeu - the graphical load did appear irrelevant to my systems performance. However, I do consider the dddac to be much better than an imic, therefore until I have tried a dac comparable to the dddac I do not dispute Klaus' findings.
I have another motive to use the commandline on a older machine - even xfce made the system sluggish at points, with the command line only it is much faster. Fluxbox is also an improvement over xfce.
Look forward to hearing the competition results.
Ross
brutus
Hello all,
I just registered to this forum after reading this topic a few times. I am using Linux for audio playback since januari this year and with a lot of help from Soundcheck I recently managed to get 'brutus' playing.
So what can I add to the Wiki that is already available. Not much, I followed all the steps mentioned in there, and I am even ahead since I do already use brutus.
I started to use a laptop for music playback after very enthousiatic stories on http://forum.bd-design.nl/forum.php. At that time I had a Theta DS Pre and a Theta Data Basic for CD playback. These were quite expensive and I liked them very much.
My start with computer playback was Windows XP with Foobar. The digital data are send through the USB port to a TwinDac (www.twindac.com). To my amazement this sounded a lot better than my Thetas.
After that comes the story with Linux and step by step the sounds gets better, i.e. closer to the acoustic original. It took a lot of time and a small investment (additional memory and hard disk) but the result is truly amazing. I recently dared to use Foobar again and I could not imagine that I liked this sound just over half a year ago.
I heard brutus up to now only on my own equipment but soon I will hear it on a system with better speakers and amplifiers so am very curious. But now I am already amazed how good CDs can sound and I wonder if we are already close to or at the limit of what can be obtained.
Kind regards,
Eddie
Hello all,
I just registered to this forum after reading this topic a few times. I am using Linux for audio playback since januari this year and with a lot of help from Soundcheck I recently managed to get 'brutus' playing.
So what can I add to the Wiki that is already available. Not much, I followed all the steps mentioned in there, and I am even ahead since I do already use brutus.
I started to use a laptop for music playback after very enthousiatic stories on http://forum.bd-design.nl/forum.php. At that time I had a Theta DS Pre and a Theta Data Basic for CD playback. These were quite expensive and I liked them very much.
My start with computer playback was Windows XP with Foobar. The digital data are send through the USB port to a TwinDac (www.twindac.com). To my amazement this sounded a lot better than my Thetas.
After that comes the story with Linux and step by step the sounds gets better, i.e. closer to the acoustic original. It took a lot of time and a small investment (additional memory and hard disk) but the result is truly amazing. I recently dared to use Foobar again and I could not imagine that I liked this sound just over half a year ago.
I heard brutus up to now only on my own equipment but soon I will hear it on a system with better speakers and amplifiers so am very curious. But now I am already amazed how good CDs can sound and I wonder if we are already close to or at the limit of what can be obtained.
Kind regards,
Eddie
Hi i have also been struggling around with linuxaudio for some years now.
I have build and used a via me6000 (600mhz fanless) motherboard, connected to DDDAC full setup (usb+tent clock, s/pdif reciver+tent clock, and 1 12xtda1543 tower) The system was running with a touchscreen + realmagic remote also.
Well i had quit the system, cause i didn't wanted all that stuff in my livingroom.
I played music via freevo (using mplayer hw:1) and no dmix interfearring!, the music was amazing (flac encoded through EAC running under linux with wine), this was working great. Haven't really tested other setup's.
I am really interessted in linuxaudio, and i am very glad to hear, that it beat's any MS$ setup ... Very good work soundcheck!
My personal opinium, is that a mediacenter isen't the way to go, it's much better to use a laptop (my next project ) to control, what you want to hear... everything is easier (also for other people) with a laptop. Or even better is the way Peufeu is going i think. with a streaming server playing music through a dac, but this is also possible with a laptop, taking data from a lowend server in youre closet
--- Controlling volume should be done at your'e Preamp, just as from your'e other equipment for easy integration --> period.
I see that with the prices at HD's today i maybee made a mistake converting my 150cd's into FLAC, but i really don't think the maybe 5-10% cpuload extracting them really matters ? OR does it ???
Well i would like any script, to automatically load the cd or even better a single number at a time into the ramdisk ? (I only have 1gb ram) ... so it's not possible for me to load more than max 512mb. Is this possible Soundcheck ?
Jesper.
I have build and used a via me6000 (600mhz fanless) motherboard, connected to DDDAC full setup (usb+tent clock, s/pdif reciver+tent clock, and 1 12xtda1543 tower) The system was running with a touchscreen + realmagic remote also.
Well i had quit the system, cause i didn't wanted all that stuff in my livingroom.
I played music via freevo (using mplayer hw:1) and no dmix interfearring!, the music was amazing (flac encoded through EAC running under linux with wine), this was working great. Haven't really tested other setup's.
I am really interessted in linuxaudio, and i am very glad to hear, that it beat's any MS$ setup
... Very good work soundcheck!My personal opinium, is that a mediacenter isen't the way to go, it's much better to use a laptop (my next project
) to control, what you want to hear... everything is easier (also for other people) with a laptop. Or even better is the way Peufeu is going i think. with a streaming server playing music through a dac, but this is also possible with a laptop, taking data from a lowend server in youre closet --- Controlling volume should be done at your'e Preamp, just as from your'e other equipment for easy integration --> period.
I see that with the prices at HD's today i maybee made a mistake converting my 150cd's into FLAC, but i really don't think the maybe 5-10% cpuload extracting them really matters ? OR does it ???
Well i would like any script, to automatically load the cd or even better a single number at a time into the ramdisk ? (I only have 1gb ram)
... so it's not possible for me to load more than max 512mb. Is this possible Soundcheck ?Jesper.
lykkedk said:IOr even better is the way Peufeu is going i think. with a streaming server playing music through a dac, but this is also possible with a laptop, taking data from a lowend server in youre closet
--- Controlling volume should be done at your'e Preamp, just as from your'e other equipment for easy integration --> period.
I see that with the prices at HD's today i maybee made a mistake converting my 150cd's into FLAC, but i really don't think the maybe 5-10% cpuload extracting them really matters ? OR does it ???
Well i would like any script, to automatically load the cd or even better a single number at a time into the ramdisk ? (I only have 1gb ram)
Jesper.
Hi Jesper.
Peufeu's approach got IMO some issues:
1.The FGPA is IMO much to weak.
2. He got several instances in the whole chain.
3. Ethernet streaming, especially on a 100MBIT is eating resources.
4. He is not able to do full-file-RAM-buffering. ( Loading a .wav- track to RAM would take him at least 20s, if he would have the
resources available.
The only advantage I see is the clocking issue. But I think there are more easy solutions to achieve this.
However:
1GB of RAM is OK for RAM-playback.
My player script supports the features you're looking at - playing
from RAM on a track by track basis.
Klaus
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.