28 Mar
2011
28 Mar
'11
11:15 p.m.
Hi Peder. Thanks for your assistance. Info below:
What CPU do you have (grep "model name"
/proc/cpuinfo)?
AMD Phenom(tm) 9750 Quad-Core Processor
Have you made the limits.conf hack to enable realtime
and are you
I'm afraid I have no idea. I have a real-time kernel and I have it set to on
when I use firewire, but off when I use alsa. All music production (when I
get the one-off xruns) happens only when using firewire, and thus with
realtime on.
running jack in realtime (with say -P70)?
Yes I have jack in realtime, on exactly P70, but ONLY in firewire mode. Alsa
mode is set to default priority and non-realtime. I have QjackCtl setup so
that it launches and starts in Alsa-mode when I boot my computer. When I do
music stuff and turn my firewire device on I switch to the "firewire" preset
which is realtime P70.
Are you running jack with ?ber-low latency; what's
the frames/period
setting?
In firewire mode I have frames/period setting of 256 with buffer on 3 = a
latency of 17.4 msec. Alsa mode is much higher on 1024 and around 70 msec
(but I believe that should be irrelevant as I don't do any audio/midi jack
connection stuff in Alsa mode.
-Dan
30 Mar
30 Mar
8:18 a.m.
New subject: xrun problems. (Was Re: Introduction to Ardour 3.0 MIDI)
Quoting Dan Capp <reflectremain(a)gmail.com>om>:
Hi Peder. Thanks for your assistance. Info below:
What CPU do you have (grep "model name"
/proc/cpuinfo)?
AMD Phenom(tm) 9750 Quad-Core Processor
Have you made the limits.conf hack to enable
realtime and are you
I'm afraid I have no idea. I have a real-time kernel and I have it set to on
when I use firewire, Yes I have jack in realtime, on exactly P70, but ONLY
in firewire mode. Alsa
mode is set to default priority and non-realtime.
In firewire mode I have frames/period setting of 256
with buffer on 3 = a
latency of 17.4 msec.
Hmm, with those specs I don't have a clue. I can't see any direct
reason you should get xruns.
But since you don't get them while recording they're completely
harmless and if the esthetics bother you you can clear them by
clicking the Messages button in QjackCtl and click Reset on the Status
tab.
1:02 p.m.
New subject: xrun problems. (Was Re: Introduction to Ardour 3.0 MIDI)
On Wed, Mar 30, 2011 at 2:18 AM, Peder Hedlund <peder(a)musikhuset.org> wrote:
Hmm, with those specs I don't have a clue. I
can't see any direct reason you
should get xruns.
i have a 6 core athlon phenom II X6 1090T (3.6GHz) system. i get xruns
"regularly". i don't know why.
as if that will make anyone feel better.
1:09 p.m.
New subject: xrun problems. (Was Re: Introduction to Ardour 3.0 MIDI)
On 30 March 2011 13:02, Paul Davis <paul(a)linuxaudiosystems.com> wrote:
i have a 6 core athlon phenom II X6 1090T (3.6GHz)
system. i get xruns
"regularly". i don't know why.
as if that will make anyone feel better.
It warms my cold heart Paul ;)
2:17 p.m.
New subject: xrun problems. (Was Re: Introduction to Ardour 3.0 MIDI)
Paul Davis wrote:
i have a 6 core athlon phenom II X6 1090T (3.6GHz)
system. i get xruns
"regularly". i don't know why.
Modern power management mechanisms are quite aggressive, but increase
latency because of all the time needed to wake the hardware up again.
Try disabling C1E in the BIOS, and/or try disabling CPU frequency
scaling.
If the xruns do happen in regular time intervals, your problem might be
the BIOS which does something in SMIs (system management interrupts).
In that case, your only choices are to disable that BIOS function (if
possible; AFAIK only some IBM servers have that option:
<http://publib.boulder.ibm.com/infocenter/lnxinfo/v3r0m0/topic/liaai/realtime/liaairtsmi.htm>),
or to get a different mainboard.
Regards,
Clemens
2:32 p.m.
New subject: xrun problems. (Was Re: Introduction to Ardour 3.0 MIDI)
On Wed, Mar 30, 2011 at 8:18 AM, Clemens Ladisch <clemens(a)ladisch.de> wrote:
Paul Davis wrote:
already off.
i have a 6 core athlon phenom II X6 1090T
(3.6GHz) system. i get xruns
"regularly". i don't know why.
Modern power management mechanisms are quite aggressive, but increase
latency because of all the time needed to wake the hardware up again.
Try disabling C1E in the BIOS, and/or try disabling CPU frequency
scaling. If the xruns do happen in regular time intervals, your
problem might be
i put "regular" in quotes to avoid the implication of "periodic".
they
are not periodic, but they happen randomly but with enough frequency
that anyone would notice them. they occur at close to zero system
load.
i tried an RT kernel from CCRMA and it hosed the video driver (nouveau).
31 Mar
31 Mar
9:45 p.m.
New subject: xrun problems. (Was Re: Introduction to Ardour 3.0 MIDI)
On 03/30/2011 05:32 AM, Paul Davis wrote:
On Wed, Mar 30, 2011 at 8:18 AM, Clemens
Ladisch<clemens(a)ladisch.de> wrote:
Which is the non-rt that works for you with nouveau?
rt-land is stuck for now in ancient 2.6.33.x kernels...
-- Fernando
Paul Davis wrote:
already off.
i have a 6 core athlon phenom II X6 1090T
(3.6GHz) system. i get xruns
"regularly". i don't know why.
Modern power management mechanisms are quite aggressive, but increase
latency because of all the time needed to wake the hardware up again.
Try disabling C1E in the BIOS, and/or try disabling CPU frequency
scaling. If the xruns do happen in regular time intervals,
your problem might be
i put "regular" in quotes to avoid the implication of "periodic".
they
are not periodic, but they happen randomly but with enough frequency
that anyone would notice them. they occur at close to zero system
load.
i tried an RT kernel from CCRMA and it hosed the video driver (nouveau).
10:38 p.m.
New subject: xrun problems. (Was Re: Introduction to Ardour 3.0 MIDI)
On Thu, Mar 31, 2011 at 3:45 PM, Fernando Lopez-Lezcano
<nando(a)ccrma.stanford.edu> wrote:
Linux sextet 2.6.34.7-66.fc13.x86_64 #1 SMP Wed Dec 15 07:04:30 UTC
2010 x86_64 x86_64 x86_64 GNU/Linux
On 03/30/2011 05:32 AM, Paul Davis wrote:
>
> On Wed, Mar 30, 2011 at 8:18 AM, Clemens Ladisch<clemens(a)ladisch.de>
> wrote:
>>
>> Paul Davis wrote:
i tried an RT
kernel from CCRMA and it hosed the video driver (nouveau).
Which is the non-rt that works for you with nouveau?
rt-land is stuck for now in ancient 2.6.33.x kernels...
4 Apr
4 Apr
9:03 p.m.
New subject: xrun problems. (Was Re: Introduction to Ardour 3.0 MIDI)
On 03/31/2011 01:38 PM, Paul Davis wrote:
On Thu, Mar 31, 2011 at 3:45 PM, Fernando
Lopez-Lezcano
<nando(a)ccrma.stanford.edu> wrote:
Linux sextet 2.6.34.7-66.fc13.x86_64 #1 SMP Wed Dec 15 07:04:30 UTC
2010 x86_64 x86_64 x86_64 GNU/Linux
Ah, ok, so it is not _that_ new. There is a recent branch of rt for
2.6.34 and I do have an rt kernel based on the 68 Fedora Koji build
(kernel-rtPAE-2.6.34.8-68.rt0.3.fc13.ccrma.i686). I have not released it
as I found problems with suspend that I have had no time to even report
them, much less fix them - otherwise seems to work fine. Drop me a line
if you would like to test drive it...
-- Fernando
On 03/30/2011 05:32 AM, Paul Davis wrote:
>
> On Wed, Mar 30, 2011 at 8:18 AM, Clemens Ladisch<clemens(a)ladisch.de>
> wrote:
>>
>> Paul Davis wrote:
i tried
an RT kernel from CCRMA and it hosed the video driver (nouveau).
Which is the non-rt that works for you with nouveau?
rt-land is stuck for now in ancient 2.6.33.x kernels... 
30 Mar
30 Mar
1:19 p.m.
New subject: xrun problems. (Was Re: Introduction to Ardour 3.0 MIDI)
Am 28.03.2011 23:15, schrieb Dan Capp:
the file /etc/security/limits.conf should have the following lines at
its end:
@audio - rtprio 86
@audio - memlock unlimited
in Ubuntu there is a file called "audio.conf" in a folder named
/etc/security/limits.d . In Fedora the file is called 99jack.conf.
All with the same lines. You need to be member in the group "audio" (or
"jackuser" under Fedora)
Hi Peder. Thanks for your assistance. Info below:
I had a Presonus Firebox running with Frames/Periods 64/3 on a ancient
Pentium4 with zero xruns. The very same Firebox spat out xruns by the
score if running it on a Lenovo Laptop with Intel Dualcore with 256/3.
Firewire Interfaces tend to be very picky regarding the chipset oft the
FW-Controller:
http://ffado.org/?q=node/251
and they amok, if there is something on the bus blocking them.
What CPU do you have (grep "model name"
/proc/cpuinfo)?
AMD Phenom(tm) 9750 Quad-Core Processor Have you made the limits.conf hack to enable
realtime and are you
I'm afraid I have no idea. I have a real-time kernel and I have it set to on
when I use firewire, but off when I use alsa. All music production (when I
get the one-off xruns) happens only when using firewire, and thus with
realtime on.
On a system like yours any Firewire-Interface should do much better (see
above).
Ever tried a Live-DVD like AVLinux or Pure:Dyne?
running jack in realtime (with say -P70)?
Yes I have jack in realtime, on exactly P70, but ONLY in firewire mode. Alsa
mode is set to default priority and non-realtime. I have QjackCtl setup so
that it launches and starts in Alsa-mode when I boot my computer. When I do
music stuff and turn my firewire device on I switch to the "firewire" preset
which is realtime P70.
Are you running jack with ?ber-low latency;
what's the frames/period
setting?
In firewire mode I have frames/period setting of 256 with buffer on 3 = a
latency of 17.4 msec. Alsa mode is much higher on 1024 and around 70 msec
(but I believe that should be irrelevant as I don't do any audio/midi jack
connection stuff in Alsa mode.
-Dan
_______________________________________________
Linux-audio-user mailing list
Linux-audio-user(a)lists.linuxaudio.org
http://lists.linuxaudio.org/listinfo/linux-audio-user
4:51 p.m.
New subject: xrun problems. (Was Re: Introduction to Ardour 3.0 MIDI)
i guess you guys know the realtimeconfigquickscan script?
http://code.google.com/p/realtimeconfigquickscan/
if not, it could help out to solve some problems
Greets
5 p.m.
New subject: xrun problems. (Was Re: Introduction to Ardour 3.0 MIDI)
Hi,
I've done all the optimizations recommended by the
realtimeconfigquickscan script detailed here below.
I followed all the recommendations included in the jackaudio.org FAQ.
I've recompiled my own rt kernel from peguntronix and kernel.org also.
I'm also get xruns with the system without load and I've observed that
just disabling the wifi network helps a lot and no more xruns are
observed.
I'm using ubuntu 9.10 with Jack1 and a cheap (but decent) Behringer
UCA202 audio i/f.
Bye
Diego
2011/3/30 Fabio <capoeirista(a)arcor.de>de>:
i guess you guys know the realtimeconfigquickscan
script?
http://code.google.com/p/realtimeconfigquickscan/
if not, it could help out to solve some problems
Greets
_______________________________________________
Linux-audio-user mailing list
Linux-audio-user(a)lists.linuxaudio.org
http://lists.linuxaudio.org/listinfo/linux-audio-user
5:11 p.m.
New subject: xrun problems. (Was Re: Introduction to Ardour 3.0 MIDI)
On Wed, Mar 30, 2011 at 10:00 AM, Diego Simak <diego.simak(a)gmail.com> wrote:
Hi,
I've done all the optimizations recommended by the
realtimeconfigquickscan script detailed here below.
I followed all the recommendations included in the jackaudio.org FAQ.
I've recompiled my own rt kernel from peguntronix and kernel.org also.
I'm also get xruns with the system without load and I've observed that
just disabling the wifi network helps a lot and no more xruns are
observed.
I'm using ubuntu 9.10 with Jack1 and a cheap (but decent) Behringer
UCA202 audio i/f.
Bye
Diego
Do you know if your wifi card is sharing an interrupt with your audio
interface?
On my laptop the nvidia card was sharing an interrupt with the firewire
interface. After enabling MSI in the nvidia driver my xruns became
non-existent. Maybe you have this option with your wifi module.
--Steve
6:11 p.m.
New subject: xrun problems. (Was Re: Introduction to Ardour 3.0 MIDI)
I found that I could run 0.363 ms latency, quite easily. No dropouts, if
just running renoise.
I did compile a kernel where I turned off all that I could though. What I
remember offhand - maximum level of preemption (no need for rt kernel).
Preemptible RCU,
SLAB (doesn't do as much stuff as slub, and probably therefore has lower
latencies)
I didn't find enabling tickless to be a problem.
250hz timer.
Turning off IOMMU's, and similar things. You don't need them on boxes that
has below 4GB RAM. IOMMU adds to latency.
Disabling all debugging, hacking, securtity, and other stuff you don't
need.
Also you need to make sure that realtime-threads get the highest priority.
I also had to modify rtprio to 99 and nice to -20, in
/etc/security/limits.d/audio.conf on ubuntu natty.
This is my full local config.
http://pastebin.com/JCie5RRN
I'm on a core2duo, also using a nvidia gtx280, without problems. 0.363 ms
latency was with renoise, jackd and konnekt24d.
The kernel is 2.6.36-zen0
http://downloads.zen-kernel.org/snapshots/zen-stable/linux-2.6-zen.v2.6.36-…
I use this because it has some low latency patches, from 38 and 37 on top
of 36, which has the old firewire stack.
Peace,
Uwaysi.
9:22 p.m.
New subject: xrun problems. (Was Re: Introduction to Ardour 3.0 MIDI)
My experience is the following :
I had a serious xruns problem using both the internal sound card of my
laptop and my M-Audio MobilePre USB sound card.
After trying a lot of things, the only config that worked for me was to set
my rtprio at priority of the real-time clock minus 11 (so that the jack
watchdog is just under the real-time clock).
These are rather low values : priority is 50 for the rtc and 39 for rtprio
in my limits.conf
but since i set up this configuration, everything runs smoothly at just
under 10 ms using 3 instances of yoshimi, hydrogen, seq24 ardour and
eventually a gx_head with my guitar as input. No xruns at all !!
I used "ps -eo rtprio,cmd" to check the rtc value with my realtime kernel.
hope it can be of any help...
jy
2011/3/30 Uwaysi Bin Kareem <uwaysi.bin.kareem(a)paradoxuncreated.com>
I found that I could run 0.363 ms latency, quite
easily. No dropouts, if
just running renoise.
I did compile a kernel where I turned off all that I could though. What I
remember offhand - maximum level of preemption (no need for rt kernel).
Preemptible RCU,
SLAB (doesn't do as much stuff as slub, and probably therefore has lower
latencies)
I didn't find enabling tickless to be a problem.
250hz timer.
Turning off IOMMU's, and similar things. You don't need them on boxes that
has below 4GB RAM. IOMMU adds to latency.
Disabling all debugging, hacking, securtity, and other stuff you don't
need.
Also you need to make sure that realtime-threads get the highest priority.
I also had to modify rtprio to 99 and nice to -20, in
/etc/security/limits.d/audio.conf on ubuntu natty.
This is my full local config.
http://pastebin.com/JCie5RRN
I'm on a core2duo, also using a nvidia gtx280, without problems. 0.363 ms
latency was with renoise, jackd and konnekt24d.
The kernel is 2.6.36-zen0
http://downloads.zen-kernel.org/snapshots/zen-stable/linux-2.6-zen.v2.6.36-…
I use this because it has some low latency patches, from 38 and 37 on top
of 36, which has the old firewire stack.
Peace,
Uwaysi.
_______________________________________________
Linux-audio-user mailing list
Linux-audio-user(a)lists.linuxaudio.org
http://lists.linuxaudio.org/listinfo/linux-audio-user
31 Mar
31 Mar
10:45 a.m.
New subject: xrun problems. (Was Re: Introduction to Ardour 3.0 MIDI)
I have a UCA202, also. Make sure it's the *only* device attached to
whatever USB controller you've hooked it to. USB sucks when multiple
devices share a controller! On my setup, it wouldn't work correctly at
all until I put it on its own USB port.
On my laptop, even with the wireless antenna turned off, the wireless
driver period pops up and scans for networks. Of course, it finds none,
but while it's interacting with the wireless hardware, it causes xruns.
IIRC, some folk on the list said to blacklist the wireless module so it
doesn't even get loaded?
I've also heard that the binary NVidia drivers do not cooperate with RT
work.
I've glad you were able to make a Pengutronix kernel work. I couldn't
make their kernel work at all. So far, the RT kernel in Aptosid seems to
be working very well on my desktop PC, so I'm planning to try it on my
laptop.
Diego Simak wrote:
Hi,
I've done all the optimizations recommended by the
realtimeconfigquickscan script detailed here below.
I followed all the recommendations included in the jackaudio.org FAQ.
I've recompiled my own rt kernel from peguntronix and kernel.org also.
I'm also get xruns with the system without load and I've observed that
just disabling the wifi network helps a lot and no more xruns are
observed.
I'm using ubuntu 9.10 with Jack1 and a cheap (but decent) Behringer
UCA202 audio i/f.
Bye
Diego
2011/3/30 Fabio <capoeirista(a)arcor.de>de>:
> i guess you guys know the realtimeconfigquickscan script?
> http://code.google.com/p/realtimeconfigquickscan/
>
> if not, it could help out to solve some problems
>
>
> Greets
--
David
gnome(a)hawaii.rr.com
authenticity, honesty, community
3:06 p.m.
New subject: xrun problems. (Was Re: Introduction to Ardour 3.0 MIDI)
2011/3/31 david <gnome(a)hawaii.rr.com>om>:
I have a UCA202, also. Make sure it's the *only*
device attached to whatever
USB controller you've hooked it to. USB sucks when multiple devices share a
controller! On my setup, it wouldn't work correctly at all until I put it on
its own USB port.
I have used a mouse with another USB port, don't really know if it
uses the same controller, but good to know that.
On my laptop, even with the wireless antenna turned off, the wireless driver
period pops up and scans for networks. Of course, it finds none, but while
it's interacting with the wireless hardware, it causes xruns. IIRC, some
folk on the list said to blacklist the wireless module so it doesn't even
get loaded?
I'm using Gnome what I'm doing is todeselect the "Enable Networking"
in the nm-applet and also turn off the wireless RF.
I still need to learn more about the usb audio i/f and see if the wifi
device support MSi that would help as was said before.
I've also heard that the binary NVidia drivers do
not cooperate with RT
work.
I have a ATI Radeon driver installed, I will try to remove this driver
and see how it works.
I've glad you were able to make a Pengutronix
kernel work. I couldn't make
their kernel work at all. So far, the RT kernel in Aptosid seems to be
working very well on my desktop PC, so I'm planning to try it on my laptop.
I just followed the instructions described here:
http://www.pengutronix.de/software/linux-rt/debian_en.html
What's the Aptosid RT kernel version?
Thank you very much.
Diego
Diego Simak wrote:
Hi,
I've done all the optimizations recommended by the
realtimeconfigquickscan script detailed here below.
I followed all the recommendations included in the jackaudio.org FAQ.
I've recompiled my own rt kernel from peguntronix and kernel.org also.
I'm also get xruns with the system without load and I've observed that
just disabling the wifi network helps a lot and no more xruns are
observed.
I'm using ubuntu 9.10 with Jack1 and a cheap (but decent) Behringer
UCA202 audio i/f.
Bye
Diego
2011/3/30 Fabio <capoeirista(a)arcor.de>de>:
>
> i guess you guys know the realtimeconfigquickscan script?
> http://code.google.com/p/realtimeconfigquickscan/
>
> if not, it could help out to solve some problems
>
>
> Greets
--
David
gnome(a)hawaii.rr.com
authenticity, honesty, community
_______________________________________________
Linux-audio-user mailing list
Linux-audio-user(a)lists.linuxaudio.org
http://lists.linuxaudio.org/listinfo/linux-audio-user
1 Apr
1 Apr
7:15 a.m.
New subject: xrun problems. (Was Re: Introduction to Ardour 3.0 MIDI)
Diego Simak wrote:
I think you actually have to blacklist the wireless driver to prevent it
from being loaded by the kernel.
I have NVidia GeForce 6150SE nForce 430 (integrated chipset), I think
aptosid is using the open-source driver (not the binary blob driver). No
need for 3D acceleration.
2.6.37. They only supply RT kernels.
--
David
gnome(a)hawaii.rr.com
authenticity, honesty, community
2011/3/31 david <gnome(a)hawaii.rr.com>om>:
My little laptop has a lot of USB things hooked to it!
I have a UCA202, also. Make sure it's the
*only* device attached to whatever
USB controller you've hooked it to. USB sucks when multiple devices share a
controller! On my setup, it wouldn't work correctly at all until I put it on
its own USB port.
I have used a mouse with another USB port, don't really know if it
uses the same controller, but good to know that. On my laptop,
even with the wireless antenna turned off, the wireless driver
period pops up and scans for networks. Of course, it finds none, but while
it's interacting with the wireless hardware, it causes xruns. IIRC, some
folk on the list said to blacklist the wireless module so it doesn't even
get loaded?
I'm using Gnome what I'm doing is todeselect the "Enable Networking"
in the nm-applet and also turn off the wireless RF.
I still need to learn more about the usb audio i/f and see if the wifi
device support MSi that would help as was said before. I've also
heard that the binary NVidia drivers do not cooperate with RT
work.
I have a ATI Radeon driver installed, I will try to remove this driver
and see how it works. I've glad
you were able to make a Pengutronix kernel work. I couldn't make
their kernel work at all. So far, the RT kernel in Aptosid seems to be
working very well on my desktop PC, so I'm planning to try it on my laptop.
I just followed the instructions described here:
http://www.pengutronix.de/software/linux-rt/debian_en.html
What's the Aptosid RT kernel version? 
10:37 a.m.
New subject: xrun problems. (Was Re: Introduction to Ardour 3.0 MIDI)
On 04/01/2011 07:15 AM, david wrote:
2.6.37. They only supply RT kernels.
Thomas Gleixner has announced the next release of the rt patch will be
based on 2.6.37 but is there already something available?
Best,
Jeremy
9:31 p.m.
New subject: xrun problems. (Was Re: Introduction to Ardour 3.0 MIDI)
Jeremy Jongepier wrote:
On 04/01/2011 07:15 AM, david wrote:
Check aptosid: http://aptosid.com.
--
David
gnome(a)hawaii.rr.com
authenticity, honesty, community
2.6.37. They only supply RT kernels.
Thomas Gleixner has announced the next release of the rt patch will be
based on 2.6.37 but is there already something available?
2 Apr
2 Apr
12:34 p.m.
New subject: xrun problems. (Was Re: Introduction to Ardour 3.0 MIDI)
On 04/01/2011 09:31 PM, david wrote:
Jeremy Jongepier wrote:
Hello David,
I've checked. No real-time kernel afaics.
Best,
Jeremy
On 04/01/2011 07:15 AM, david wrote:
Check aptosid: http://aptosid.com.
2.6.37. They only supply RT kernels.
Thomas Gleixner has announced the next release of the rt patch will be
based on 2.6.37 but is there already something available? 
8:13 p.m.
New subject: xrun problems. (Was Re: Introduction to Ardour 3.0 MIDI)
Hi all,
I don't know if this is the most correct procedure, but I got a
latency here of 0.7ms (not 7 or 70, 0.7!) using schedtool. I start
JACK with "schedtool -I -e jackd" and the latency droped from 6ms to
0.7ms.
I'm using a custom 2.6.38.2-bfs kernel on a core i7 thinkpad laptop.
I did not measure it scientifically, but I don't hear any xruns, and
qjackctl do not report any problems.
Best,
Bernardo
10:16 p.m.
New subject: xrun problems. (Was Re: Introduction to Ardour 3.0 MIDI)
On Sat, Apr 2, 2011 at 2:13 PM, Bernardo Barros
<bernardobarros2(a)gmail.com> wrote:
Hi all,
I don't know if this is the most correct procedure, but I got a
latency here of 0.7ms (not 7 or 70, 0.7!) using schedtool. I start
JACK with "schedtool -I -e jackd" and the latency droped from 6ms to
0.7ms.
what frames/period and periods/buffer settings did you use with JACK?
3 Apr
3 Apr
12:31 a.m.
New subject: xrun problems. (Was Re: Introduction to Ardour 3.0 MIDI)
2011/4/2, Paul Davis <paul(a)linuxaudiosystems.com>om>:
what frames/period and periods/buffer settings did you
use with JACK?
I was changin the arguments to see how far it would handle, I got
frames/period of 16 and periods/buffer of 2 at 44100.
12:52 a.m.
New subject: xrun problems. (Was Re: Introduction to Ardour 3.0 MIDI)
I would be very embarrassed if I'm fooling myself, but from what I
read the information provided by qjackctl, is this really the latency
of the audio system as low as 0.08ms with frames/period of 2! :S
What I did was actually only replace the argument " ServerPath " by
"schedtool -I -e jackd"
No human can perceive the diference between latencies of 8 or 1ms, but
if this is acurate (I have my doubts) it tells that audio is having a
extremily very high priority on the system at least.
2011/4/2, Bernardo Barros <bernardobarros2(a)gmail.com>om>:
2011/4/2, Paul Davis
<paul(a)linuxaudiosystems.com>om>:
what frames/period and periods/buffer settings
did you use with JACK?
I was changin the arguments to see how far it would handle, I got
frames/period of 16 and periods/buffer of 2 at 44100.
1:05 a.m.
New subject: xrun problems. (Was Re: Introduction to Ardour 3.0 MIDI)
On 04/03/2011 12:52 AM, Bernardo Barros wrote:
I would be very embarrassed if I'm fooling myself,
but from what I
read the information provided by qjackctl, is this really the latency
of the audio system as low as 0.08ms with frames/period of 2! :S
must be a typo: 2*16/ 44100Hz = 0.725ms = ~0.8ms
Measure it with 'jack_delay' [1]. The number given in qjackctl's setup
dialog is the nominal latency; and the time between two jack cycles.
Usually there's additional latency due to [PCI|USB] bus overhead,
sound-card & ALSA buffering, DAC/ADCs, etc. - around 1 ms round-trip for
PCI devices.
http://wiki.linuxaudio.org/wiki/jack_latency_tests
have fun,
robin
[1] http://kokkinizita.linuxaudio.org/linuxaudio/index.html
1:16 a.m.
New subject: xrun problems. (Was Re: Introduction to Ardour 3.0 MIDI)
No, with frame/period of 2 it also *seems* to work.
I tried jack_delay, it returns me this whatever frames/period I try.
What that means?
capture latency = 1024
playback_latency = 1024
Signal below threshold...
Signal below threshold...
Signal below threshold...
Signal below threshold...
Signal below threshold...
Signal below threshold...
2011/4/2, Robin Gareus <robin(a)gareus.org>rg>:
On 04/03/2011 12:52 AM, Bernardo Barros wrote:
I would be very embarrassed if I'm fooling
myself, but from what I
read the information provided by qjackctl, is this really the latency
of the audio system as low as 0.08ms with frames/period of 2! :S
must be a typo: 2*16/ 44100Hz = 0.725ms = ~0.8ms
Measure it with 'jack_delay' [1]. The number given in qjackctl's setup
dialog is the nominal latency; and the time between two jack cycles.
Usually there's additional latency due to [PCI|USB] bus overhead,
sound-card & ALSA buffering, DAC/ADCs, etc. - around 1 ms round-trip for
PCI devices.
http://wiki.linuxaudio.org/wiki/jack_latency_tests
have fun,
robin
[1] http://kokkinizita.linuxaudio.org/linuxaudio/index.html
1:51 a.m.
New subject: xrun problems. (Was Re: Introduction to Ardour 3.0 MIDI)
On Sat, Apr 2, 2011 at 7:16 PM, Bernardo Barros
<bernardobarros2(a)gmail.com> wrote:
No, with frame/period of 2 it also *seems* to work.
I tried jack_delay, it returns me this whatever frames/period I try.
What that means?
you must connect the output to the input. its normal to do this with a
physical cable that connects output to input on the soundcard.
2:03 a.m.
New subject: xrun problems. (Was Re: Introduction to Ardour 3.0 MIDI)
2011/4/2 Paul Davis <paul(a)linuxaudiosystems.com>
you must connect the output to the input. its normal to do this with a
physical cable that connects output to input on the soundcard.
(sorry)
starting as root it works... hum.. Must be some misconfiguration somewhere
else then...
Connecting the cables etc:
64.000 frames 1.451 ms
2 Apr
2 Apr
10:39 p.m.
New subject: xrun problems. (Was Re: Introduction to Ardour 3.0 MIDI)
Jeremy Jongepier wrote:
On 04/01/2011 09:31 PM, david wrote:
That's what it reports on my installation here.
--
David
gnome(a)hawaii.rr.com
authenticity, honesty, community
Jeremy Jongepier wrote:
Hello David,
I've checked. No real-time kernel afaics. On 04/01/2011 07:15 AM, david wrote:
Check aptosid: http://aptosid.com. 2.6.37. They only supply RT kernels.
Thomas Gleixner has announced the next release of the rt patch will be
based on 2.6.37 but is there already something available?
10:56 p.m.
New subject: xrun problems. (Was Re: Introduction to Ardour 3.0 MIDI)
david wrote:
Jeremy Jongepier wrote:
Sorry, I guess it's not an RT kernel (it has PREEMPT in uname -a output,
but not an RT). Although JACKD runs happily on it in RT mode, with no
xruns, at 1.33msec latency. When I go down to 0.667msec, I got an xrun
or two.
--
David
gnome(a)hawaii.rr.com
authenticity, honesty, community
On 04/01/2011 09:31 PM, david wrote:
That's what it reports on my installation here. Jeremy Jongepier wrote:
Hello David,
I've checked. No real-time kernel afaics. On 04/01/2011 07:15 AM, david wrote:
> 2.6.37. They only supply RT kernels.
Thomas Gleixner has announced the next release of the rt patch will be
based on 2.6.37 but is there already something available?
Check aptosid: http://aptosid.com.
4976
days inactive
4983
days old
linux-audio-user@lists.linuxaudio.org
30 comments
16 participants
participants (16)
-
Arve Barsnes -
Bernardo Barros -
Clemens Ladisch -
Dan Capp -
david -
Diego Simak -
Fabio -
Fernando Lopez-Lezcano -
Hartmut Noack -
Jeremy Jongepier -
jy -
Paul Davis -
Peder Hedlund -
Robin Gareus -
S C Rigler -
Uwaysi Bin Kareem