Re: brutefir
Jack unfortunately slows down the traffic. I tried Jack before.
You can not run the low buffers sizes I am currently running.
Keep in mind that Jack is also using Alsa as engine.
You got one more application in the loop.
However:
How are you operating brutefir?
* Input? file or pipe
* realtime config
* filter partitioning
* running as deamon
* dither
* volume control
* 64 bit
* processor load balancing
Perhaps you could drop me your config file.
One thing for sure: You also face the HD file cache issue I mentioned before, if you run files as input source. That's perhaps a different issue if you run stdin (pipe) as raw-input.
Cheers
houstonian said:
Jack unfortunately slows down the traffic. I tried Jack before.
You can not run the low buffers sizes I am currently running.
Keep in mind that Jack is also using Alsa as engine.
You got one more application in the loop.
However:
How are you operating brutefir?
* Input? file or pipe
* realtime config
* filter partitioning
* running as deamon
* dither
* volume control
* 64 bit
* processor load balancing
Perhaps you could drop me your config file.
One thing for sure: You also face the HD file cache issue I mentioned before, if you run files as input source. That's perhaps a different issue if you run stdin (pipe) as raw-input.
Cheers
BRUTUS is born.
Hi folks.
I finally solved the trickiest brutefir problem - the nasty start-up pulse.
BRUTUS, that's how I called my player ( a wrapper script build around brutefir)
will be made available to interested people soon.
(Though it needs to be made a bit more generic to meet all setups before I can do it.)
I am really wondering if there is anything better sounding around.
I havn't had a chance to listen to anything in the PC audio-world that's better sounding than BRUTUS.
Of course that's first of all valid for my own setup, with a USB-DDDAC directly coupled to the AMP and a tweaked RT-Linux.
Would be interesting to know, how it sounds on other setups:
Ross, are you interested?
Gotta run! Need to feed BRUTUS
Cheers
Hi folks.
I finally solved the trickiest brutefir problem - the nasty start-up pulse.
BRUTUS, that's how I called my player ( a wrapper script build around brutefir)
will be made available to interested people soon.
(Though it needs to be made a bit more generic to meet all setups before I can do it.)
I am really wondering if there is anything better sounding around.
I havn't had a chance to listen to anything in the PC audio-world that's better sounding than BRUTUS.
Of course that's first of all valid for my own setup, with a USB-DDDAC directly coupled to the AMP and a tweaked RT-Linux.
Would be interesting to know, how it sounds on other setups:
Ross, are you interested?
Gotta run! Need to feed BRUTUS
Cheers
Hi Klaus,
Really interested in your player. Particularly if it makes brutefir any easier to use. However Im having alsa support problems though so need to resolve those first (hopefully in the next week or so). Please give details when you have settled on your optimum arrangement.
I think my cyrix iii will really struggle to do the fir filters well given its poor floating point performance - may try IIR ladspa plugins through alsa direct. If I do have to stick with IIR, I am somewhat encouraged that some have reported that Frequency allocator (currently IIR only) can sound better than the DECX fir filter. Brutefir has not been compared. IIR would have much lower delay/latency than Brutefir in keeping with the real time setup.
Im also really keen to hear how you get on in using 2 usb dacs. I read a post somewhere from some one who did this, one dac for left, one for right and reported a big improvment in soundstage. However, I really think that unless both dacs use the same clock there will be issues, at least when listening over longer periods (there are a few posts in the other thread about this). Perhaps there are no issues at all if you share the 12mhz clock between 2 usb chips? Of course this would not allow the idea of running usb cables to each speaker.
If your dual usb experiments work out well I may splash out on the DDDac, usb monica or even something comercial but multichannel. Harware support is really still quite a struggle for alsa it seems and is gettin in the way at the moment.
Keep us posted.
Ross
Really interested in your player. Particularly if it makes brutefir any easier to use. However Im having alsa support problems though so need to resolve those first (hopefully in the next week or so). Please give details when you have settled on your optimum arrangement.
I think my cyrix iii will really struggle to do the fir filters well given its poor floating point performance - may try IIR ladspa plugins through alsa direct. If I do have to stick with IIR, I am somewhat encouraged that some have reported that Frequency allocator (currently IIR only) can sound better than the DECX fir filter. Brutefir has not been compared. IIR would have much lower delay/latency than Brutefir in keeping with the real time setup.
Im also really keen to hear how you get on in using 2 usb dacs. I read a post somewhere from some one who did this, one dac for left, one for right and reported a big improvment in soundstage. However, I really think that unless both dacs use the same clock there will be issues, at least when listening over longer periods (there are a few posts in the other thread about this). Perhaps there are no issues at all if you share the 12mhz clock between 2 usb chips? Of course this would not allow the idea of running usb cables to each speaker.
If your dual usb experiments work out well I may splash out on the DDDac, usb monica or even something comercial but multichannel. Harware support is really still quite a struggle for alsa it seems and is gettin in the way at the moment.
Keep us posted.
Ross
Hi Ross.
Anders Torger is telling me that the ALSA API mainly takes care about the sync
of the different channels. It shouldn't be that bad I guess.
But one thing after another. I guess the 2-channel stereo playback will be already pretty satisfying.
The 64bit volume control replacing the pre-amp is also highly recommended,
You need to try that.
I was also thinking about two DACs (left/right) each DAC covering Satellite and
Woofer. Channel separation should be actually great by doing it this way.
And you'd run just two USB-Cables to your active speakers!
Great to have even more options for the future.
Cheers
\Klaus
Anders Torger is telling me that the ALSA API mainly takes care about the sync
of the different channels. It shouldn't be that bad I guess.
But one thing after another. I guess the 2-channel stereo playback will be already pretty satisfying.
The 64bit volume control replacing the pre-amp is also highly recommended,
You need to try that.
I was also thinking about two DACs (left/right) each DAC covering Satellite and
Woofer. Channel separation should be actually great by doing it this way.
And you'd run just two USB-Cables to your active speakers!
Great to have even more options for the future.
Cheers
\Klaus
Hello Klaus,
I've carefully read your thread along from the time it was started. The problem is that I'll be able to do some experiments only after the end of the month (it was a busy year).
I'll probably post here if I'll have problems, but at least i'll be able to offer some feedback. I plan using xubuntu with the rt patch, brutefir and drc. In the end i'll probably need using jack because i'm planing for an ethernet DAC, so I guess I'll have some more problems to solve...
Cheers!
I've carefully read your thread along from the time it was started. The problem is that I'll be able to do some experiments only after the end of the month (it was a busy year).
I'll probably post here if I'll have problems, but at least i'll be able to offer some feedback. I plan using xubuntu with the rt patch, brutefir and drc. In the end i'll probably need using jack because i'm planing for an ethernet DAC, so I guess I'll have some more problems to solve...
Cheers!
Hi folks.
After running brutus now for some days, we really got kind of bored.
"That should be it - No way!" crossed my mind.
We took the iniative to start tweaking brutefir.
I mean -- the whole system had been tweaked.
Except the brutefir sources!
The idea was to increase the task priorities on the I/O modules on
brutefir. Anders Torger put the maximum priority on the filters and not on the IO modules. That's perhaps a good idea if you process excessive filters, but just for mainly copying data, doing volume control and dithering, at this stage. That seemed to be overkill.
Tom, figured it out. THX Tom. Just 3 parameters had to be changed in the code. And the sources had to be compiled again.
Here we go again! Low level distortions clearly improved again. This was audible within 5seconds.
It is amazing to me how much the PC SW-setup still impacts the datastream.
Cheers
Klaus
After running brutus now for some days, we really got kind of bored.
"That should be it - No way!" crossed my mind.
We took the iniative to start tweaking brutefir.
I mean -- the whole system had been tweaked.
Except the brutefir sources!
The idea was to increase the task priorities on the I/O modules on
brutefir. Anders Torger put the maximum priority on the filters and not on the IO modules. That's perhaps a good idea if you process excessive filters, but just for mainly copying data, doing volume control and dithering, at this stage. That seemed to be overkill.
Tom, figured it out.
THX Tom. Just 3 parameters had to be changed in the code. And the sources had to be compiled again.Here we go again! Low level distortions clearly improved again. This was audible within 5seconds.
It is amazing to me how much the PC SW-setup still impacts the datastream.
Cheers
Klaus
Any of you guys familiar with the BEOS operating system?
I tried it when it first came out, and my first impression was that it definatly taps more power out of a PC than windows does... kinda like DOS... unfortunately drivers available at the time was few and far between...
Seeing this thread reminded me of it, and I thought I'd do some google searches...with very interesting results...
http://www.birdhouse.org/beos/bible/exc_audio.html
I tried it when it first came out, and my first impression was that it definatly taps more power out of a PC than windows does... kinda like DOS... unfortunately drivers available at the time was few and far between...
Seeing this thread reminded me of it, and I thought I'd do some google searches...with very interesting results...
http://www.birdhouse.org/beos/bible/exc_audio.html
Just to mention where we are at right now:
With our realtime-patched-Linux and the other tweaks we applied on top of our setup we're a playing in the micro-seconds-latency-league. We're talking about not more than 32 samples (32*20us) in a stabil setup. We're now working on a stabil 16 samples setup. The only problem we have here, is an unstabil process-chain-setup from time to time. (You need to try several times to start it up)! But once it's running it's running and sounding best of course.
The only OS , which might be able to challenge my current setup is RT-Linux
http://en.wikipedia.org/wiki/RTLinux , which would be even more
efficient.
I was looking at it some time ago. The big issue is, that you have
to write your own drivers.
FYI: Even OSX provides latencies of 1ms today.
Let see where Vista ends up. Players and soundcard-drivers that are programmed for supporting the exclusice mode under Vista will also end up in that region. Still they wouldn't be able to challenge the Linux setup.
And don't forget. The support and community around your OS of choice is key to success. Vista & OSX are black box OSes. No way to tweak them.
Cheers
With our realtime-patched-Linux and the other tweaks we applied on top of our setup we're a playing in the micro-seconds-latency-league.
We're talking about not more than 32 samples (32*20us) in a stabil setup. We're now working on a stabil 16 samples setup. The only problem we have here, is an unstabil process-chain-setup from time to time. (You need to try several times to start it up)! But once it's running it's running and sounding best of course. The only OS , which might be able to challenge my current setup is RT-Linux
http://en.wikipedia.org/wiki/RTLinux , which would be even more
efficient.
I was looking at it some time ago. The big issue is, that you have
to write your own drivers.
FYI: Even OSX provides latencies of 1ms today.
Let see where Vista ends up. Players and soundcard-drivers that are programmed for supporting the exclusice mode under Vista will also end up in that region. Still they wouldn't be able to challenge the Linux setup.
And don't forget. The support and community around your OS of choice is key to success. Vista & OSX are black box OSes. No way to tweak them.
Cheers
BeOS was what Final Scratch (Time coded system for mixing audio files using vinyl turntables) was developed on. The soundplay player was key to final scratch development due to its ability to play multiple files and apply different pitch to each one. I looked at it before, but ruled it out due to lack of soundcard support.
Whilst we are discussing other options...
For a usb based system I was wondering if the original xbox might be a good and cost effective way forward. Xebian, dynebolic and other media centre editions gives you a full networkable linux environment with 4 usb ports (custom connectors) on the xbox. The memory is extremely low, 64mb, but I believe it can be upgraded with some effort. There is optical out, but believe the chip resamples to 48khz so usb would be the only realistic route for high quality audio. Could interface through the TV, run without a keyboard, and at least Xebian looks like it could be tweaked with the relevant rt patches.
Anyway, just a thought, maybe some one can confirm that they have tried a decent usb dac with the xbox.
Ross
Whilst we are discussing other options...
For a usb based system I was wondering if the original xbox might be a good and cost effective way forward. Xebian, dynebolic and other media centre editions gives you a full networkable linux environment with 4 usb ports (custom connectors) on the xbox. The memory is extremely low, 64mb, but I believe it can be upgraded with some effort. There is optical out, but believe the chip resamples to 48khz so usb would be the only realistic route for high quality audio. Could interface through the TV, run without a keyboard, and at least Xebian looks like it could be tweaked with the relevant rt patches.
Anyway, just a thought, maybe some one can confirm that they have tried a decent usb dac with the xbox.
Ross
Ubuntu-Studio 7.04 available
Hi folks.
Ubuntu-Studio 7.04 is available for download
Plug'n Play Pro-Audio for Audio-Freaks, Linux Lovers and Linux Newbies!
Check it out! Highly recommended.
http://ubuntustudio.org/
The low-latency-kernel is part of it.
I guess You'll like it - Enjoy!!
Cheers
\Klaus
Hi folks.
Ubuntu-Studio 7.04 is available for download
Plug'n Play Pro-Audio for Audio-Freaks, Linux Lovers and Linux Newbies!
Check it out! Highly recommended.
http://ubuntustudio.org/
The low-latency-kernel is part of it.
I guess You'll like it - Enjoy!!
Cheers
\Klaus
Klaus,
Do you know if the ubuntu studio project now uses the low latency patches? I was unable to confirm this.
I read this interview a while ago http://oktyabr.wordpress.com/2007/01/27/ubuntu-studio-an-interview-with-project-manager-cory-kontros/
Certainly at this point the decision was to lower latency by using the the premption options within the stock ubuntu kernel and raising the timer to 1000hz.
Have you used this distribution and found it comparable to your other setup?
Ross
Do you know if the ubuntu studio project now uses the low latency patches? I was unable to confirm this.
I read this interview a while ago http://oktyabr.wordpress.com/2007/01/27/ubuntu-studio-an-interview-with-project-manager-cory-kontros/
Certainly at this point the decision was to lower latency by using the the premption options within the stock ubuntu kernel and raising the timer to 1000hz.
Have you used this distribution and found it comparable to your other setup?
Ross
There are 3 kernels available for Ubuntu Studio. Standard, Low Latency (timer set to 1000 instead of 250 etc) and in testing a Realtime Kernel (search ubuntuforums.org for the details on how to install this)....
I am running on the realtime kernel now and the difference is like night and day. It does effect the load balance a bit (I tend to get more coasters burning if I have audio running)... But I can get much lower latencies with stability.
(I am running kubuntu rather than ubuntu studio. But I do have an installation of the later and was reasonably impressed with it).
I am running on the realtime kernel now and the difference is like night and day. It does effect the load balance a bit (I tend to get more coasters burning if I have audio running)... But I can get much lower latencies with stability.
(I am running kubuntu rather than ubuntu studio. But I do have an installation of the later and was reasonably impressed with it).
Hi folks.
Ubuntu-Studio is a reasonable package as a starting point.
It'll deliver reasonable off-the-shelve performance. And the very important support network.
The low-latency kernel is in fact far away from meeting cfs or
ck performance.
NOTE: The Ingo Molnar cfs-patch will, as it seems, be integrated into mainline kernel soon , Con Kolivas on the other hand announced last weekend to stop working on his ck-patch-set!! What a pity!!!
Ubuntu-Studio won't come with my prefered ck-rt-patch ( e.g. to run at 10000Hz and some other ck related patches ).
However, if you take the whole thing serious, you have to build your own kernel anyhow to get best performance.
Working with a "standard kernel" fetched from the repositories I'd neither recommend in general nor would I regard it to be useful for getting serious audio performance on the system.
(I could post a small how-to later on!)
In general it is said that applications should be compiled on your own system-environment for best-performance ( for hardcore linux-freaks only!)
Cheers
Klaus
Ubuntu-Studio is a reasonable package as a starting point.
It'll deliver reasonable off-the-shelve performance. And the very important support network.
The low-latency kernel is in fact far away from meeting cfs or
ck performance.
NOTE: The Ingo Molnar cfs-patch will, as it seems, be integrated into mainline kernel soon , Con Kolivas on the other hand announced last weekend to stop working on his ck-patch-set!! What a pity!!!
Ubuntu-Studio won't come with my prefered ck-rt-patch ( e.g. to run at 10000Hz and some other ck related patches ).
However, if you take the whole thing serious, you have to build your own kernel anyhow to get best performance.
Working with a "standard kernel" fetched from the repositories I'd neither recommend in general nor would I regard it to be useful for getting serious audio performance on the system.
(I could post a small how-to later on!)
In general it is said that applications should be compiled on your own system-environment for best-performance ( for hardcore linux-freaks only!)
Cheers
Klaus
Great job Klaus
Hi Klaus,
You are doing a great work for the PC Audio community.
How much of a chance does a person like me, who is not
much of a computer software freak, have, to set up your system?
Would I be able to use the normal Linux drivers for my soundcard
with your Linux version?
I ask this because I modified a soundcard with four AKM 4396 on
it, that is way better than all other sources I own, including a tweaked
DDDac. I now see a chance to use a PC crossover for my horn system,
through the 8 channels of my soundcard without much of a sonic penalty.
If I can't get a Linux system to work I will try Vista which suppose to
have a much better sound than XP.
How can I email you, I may have an interesting deal?
Greets,
Klaus
Hi Klaus,
You are doing a great work for the PC Audio community.
How much of a chance does a person like me, who is not
much of a computer software freak, have, to set up your system?
Would I be able to use the normal Linux drivers for my soundcard
with your Linux version?
I ask this because I modified a soundcard with four AKM 4396 on
it, that is way better than all other sources I own, including a tweaked
DDDac. I now see a chance to use a PC crossover for my horn system,
through the 8 channels of my soundcard without much of a sonic penalty.
If I can't get a Linux system to work I will try Vista which suppose to
have a much better sound than XP.
How can I email you, I may have an interesting deal?
Greets,
Klaus
Re: Great job Klaus
Not very many people, at least over here confirmed that. I am not sure about it.
I pretty much have the feeling I am talking to myself all the time.
Quite high, with a basic understanding of computers and the willingness to learn. A basic Ubuntu-Studio as a startingpoint, can be installed by anybody on a defragmented Windows disk, with 4GB
of spare capacity.
It's IMO easier than installing MS Windows.
For the hacking part:
I am currently leading a guy offline through that process.
Starting with Ubuntu, configuring a RAMDISK, driver, kernel, asf.
It works quite well. He's pretty much impressed about the progress.
He is btw using a USB Twin-DAC !! First time somebody with a different DAC then a DDDAC confirmed the improvements.
Beside the OS related tweaks, he seems to enjoy very much the ssrc 44.1 to 48khz conversion, with the scripts I wrote for doing it!!
I had to reinstall the system the other day. This time I put
together a step by step instruction. I'll disclose this to public soon.
The idea I mentioned earlier, to build up a specific distribution, is just too time consuming and put on hold for the time being.
Of course. As long as ALSA - the Linux sound-system is supporting it.
http://www.alsa-project.org/alsa-doc/
My PC based crossover project, is idleing for quite some time, since
I still don't have the feeling that I squeezed everything out of the stereo setup. I recently made again quite some improvements by tweaking certain scheduler parameters.
Just use the forum - Email function - I just unlocked it - I wasn't aware that it was locked.
Cheers
Klaus
Radian said:
Not very many people, at least over here confirmed that. I am not sure about it.
I pretty much have the feeling I am talking to myself all the time.
Radian said:
Quite high, with a basic understanding of computers and the willingness to learn. A basic Ubuntu-Studio as a startingpoint, can be installed by anybody on a defragmented Windows disk, with 4GB
of spare capacity.
It's IMO easier than installing MS Windows.
For the hacking part:
I am currently leading a guy offline through that process.
Starting with Ubuntu, configuring a RAMDISK, driver, kernel, asf.
It works quite well. He's pretty much impressed about the progress.
He is btw using a USB Twin-DAC !! First time somebody with a different DAC then a DDDAC confirmed the improvements.
Beside the OS related tweaks, he seems to enjoy very much the ssrc 44.1 to 48khz conversion, with the scripts I wrote for doing it!!
I had to reinstall the system the other day. This time I put
together a step by step instruction. I'll disclose this to public soon.
The idea I mentioned earlier, to build up a specific distribution, is just too time consuming and put on hold for the time being.
Radian said:
Of course. As long as ALSA - the Linux sound-system is supporting it.
http://www.alsa-project.org/alsa-doc/
Radian said:
My PC based crossover project, is idleing for quite some time, since
I still don't have the feeling that I squeezed everything out of the stereo setup. I recently made again quite some improvements by tweaking certain scheduler parameters.
Radian said:
Just use the forum - Email function - I just unlocked it - I wasn't aware that it was locked.
Cheers
Klaus
Re: Re: Great job Klaus
Hi Klaus,
You are being read -- it's just all a bit way over my head to contribute anything of use.
I know my way around XP, and have tried in the past to move to Linux, but it was tough -- and I gave up each time.
So I'm still pretty much stuck where you were a while ago -- XP, J River Media Center and a DDDac.
JRMC is the hardest thing to quit for me -- I just love to be able to browse all my CDs my album thumbnails. The system sounds great and is very family friendly, and so far I haven't anything on Linux that looks anywhere near as easy to use.
However, ubuntu studio looks interesting and perhaps a good starting point for a linux noob like me... if only there was a player that could give me the features of JRMC........ and I will no longer be tempted by vista!!!
regards stefan
soundcheck said:
Not very many people, at least over here confirmed that. I am not sure about it.
I pretty much have the feeling I am talking to myself all the time.
Hi Klaus,
You are being read -- it's just all a bit way over my head to contribute anything of use.
I know my way around XP, and have tried in the past to move to Linux, but it was tough -- and I gave up each time.
So I'm still pretty much stuck where you were a while ago -- XP, J River Media Center and a DDDac.
JRMC is the hardest thing to quit for me -- I just love to be able to browse all my CDs my album thumbnails. The system sounds great and is very family friendly, and so far I haven't anything on Linux that looks anywhere near as easy to use.
However, ubuntu studio looks interesting and perhaps a good starting point for a linux noob like me... if only there was a player that could give me the features of JRMC........ and I will no longer be tempted by vista!!!
regards stefan
OT/ We really need a separate PC Audio section.
anyone interested please vote for it:
http://www.diyaudio.com/request/
/OT
anyone interested please vote for it:
http://www.diyaudio.com/request/
/OT
ssmith said:
Here's the specific link from the list:
http://www.diyaudio.com/request/vote.php?id=98
-David
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.