22 Dec
2009
22 Dec
'09
12:26 a.m.
hey list,
I've been struggling for the past couple of weeks trying to get my
netbook to perform well as a lightweight music workstation (mostly
synth work, light daw stuff). I'm getting many, many xruns, clients
getting kicked, etc. something is clearly not right. how do I begin
tracking down the source of xruns? I have an -rt kernel, a usb sound
card using -p256 -n3, running fluxbox with only wicd running to manage
my network. point is, I've tried the hunt-and-peck method and
(surprise) it hasn't done any good. how can I go about finding out
what exactly is tripping jack up, and what I can do to stop it?
(PS - I've said it before: maybe I'm expecting too much of a netbook.
could that be it?)
--
Josh Lawrence
http://www.hardbop200.com
22 Dec
22 Dec
12:45 a.m.
On 12/22/2009 03:26 PM, Josh Lawrence wrote:
hey list,
I've been struggling for the past couple of weeks trying to get my
netbook to perform well as a lightweight music workstation (mostly
synth work, light daw stuff). I'm getting many, many xruns, clients
getting kicked, etc. something is clearly not right. how do I begin
tracking down the source of xruns? I have an -rt kernel, a usb sound
card using -p256 -n3, running fluxbox with only wicd running to manage
my network. point is, I've tried the hunt-and-peck method and
(surprise) it hasn't done any good. how can I go about finding out
what exactly is tripping jack up, and what I can do to stop it?
(PS - I've said it before: maybe I'm expecting too much of a netbook.
could that be it?)
Many people have had great success running netbooks at very low latency
so your not expecting too much.
How does the netbook perform with the onboard sound card?
Are you running the usb device at 44100 or 48000? Ideally usb cards
should be run at 48000 with period sizes that divide perfectly into
48000. i.e 100, 200, 400, etc...
You can also enable/disable various kernel and alsa features to improve
performance.
Cheers.
Patrick Shirkey
Boost Hardware Ltd
2:47 a.m.
New subject: USB period size? (WAS: tracking down xruns)
On Tue, 2009-12-22 at 15:44 +1100, Patrick Shirkey wrote:
Are you running the usb device at 44100 or 48000? Ideally usb cards
should be run at 48000 with period sizes that divide perfectly into
48000. i.e 100, 200, 400, etc...
Could I know more about this, please? 48000 would be the 'generally'
used (besides 96k) for USB sound cards I understand, but what's this
about period sizes? I've always thought period sizes were of the series
8, 16, 32, 64, 128, 256...
3:29 a.m.
New subject: USB period size? (WAS: tracking down xruns)
Ng Oon-Ee wrote:
On Tue, 2009-12-22 at 15:44 +1100, Patrick Shirkey
wrote:
I did some experimenting with period sizes on my USB card (Behringer
UCA-202 running at 48k) using jackdmp on my Debian Lenny non-RT ordinary
use laptop. It was a bit better about overruns, but not much better than
the 1024 (64msec latency) I usually run here on my non-RT kernel.
Let's try it on musicbox (also non RT, but running regular jack).
Command line reports that frames must be a power of 2: 64, 512, etc. I
tried a period of 100 (worked with jackdmp). I think it's jack 0.116 on
musicbox.
--
David
gnome(a)hawaii.rr.com
authenticity, honesty, community
Are you running the usb device at 44100 or 48000?
Ideally usb cards
should be run at 48000 with period sizes that divide perfectly into
48000. i.e 100, 200, 400, etc...
Could I know more about this, please? 48000 would be the 'generally'
used (besides 96k) for USB sound cards I understand, but what's this
about period sizes? I've always thought period sizes were of the series
8, 16, 32, 64, 128, 256...
4:12 a.m.
New subject: USB period size? (WAS: tracking down xruns)
On 12/22/2009 05:47 PM, Ng Oon-Ee wrote:
On Tue, 2009-12-22 at 15:44 +1100, Patrick Shirkey
wrote:
Normally, period sizes as you describe above are the right way to go but
with usb devices the packet sizes are different to pci cards so you will
see better performance with these strange period sizes. That's not to
say the card won't work it's just that you will be transferring more
packets with period sizes that are not exact multiples of the frame rate
and putting the system under more stress.
At least that's what I understand from Takashi Iwai who wrote the first
version of the alsa usb driver.
I have personally found that with a fast enough cpu and plenty of memory
that I can run a usb-1.0 device at 64 periods and 44100 frame rate with
-n3 and get 4 channels of i/o.
Cheers.
Patrick Shirkey
Boost Hardware Ltd
Are you running the usb device at 44100 or 48000?
Ideally usb cards
should be run at 48000 with period sizes that divide perfectly into
48000. i.e 100, 200, 400, etc...
Could I know more about this, please? 48000 would be the
'generally'
used (besides 96k) for USB sound cards I understand, but what's this
about period sizes? I've always thought period sizes were of the series
8, 16, 32, 64, 128, 256...
10:31 a.m.
On Mon, Dec 21, 2009 at 10:44 PM, Patrick Shirkey
<pshirkey(a)boosthardware.com> wrote:
Are you running the usb device at 44100 or 48000?
Ideally usb cards
should be run at 48000 with period sizes that divide perfectly into
48000. i.e 100, 200, 400, etc...
I didn't know any of this, thank you!
--
Josh Lawrence
http://www.hardbop200.com
2:10 a.m.
2009/12/22 Josh Lawrence <hardbop200(a)gmail.com>om>:
hey list,
I've been struggling for the past couple of weeks trying to get my
netbook to perform well as a lightweight music workstation (mostly
synth work, light daw stuff). I'm getting many, many xruns, clients
getting kicked, etc. something is clearly not right. how do I begin
tracking down the source of xruns? I have an -rt kernel, a usb sound
card using -p256 -n3, running fluxbox with only wicd running to manage
my network. point is, I've tried the hunt-and-peck method and
(surprise) it hasn't done any good. how can I go about finding out
what exactly is tripping jack up, and what I can do to stop it?
Most probably you have an intel onboard gpu:
http://bugs.freedesktop.org/show-bug.cgi?id=23739
--
GPG/PGP ID: B42DDCAD
2:15 a.m.
2009/12/22 Ray Rashif <schivmeister(a)gmail.com>om>:
Most probably you have an intel onboard gpu:
http://bugs.freedesktop.org/show-bug.cgi?id=23739
Oops..it's http://bugs.freedesktop.org/show_bug.cgi?id=23739
--
GPG/PGP ID: B42DDCAD
10:32 a.m.
On Tue, Dec 22, 2009 at 12:14 AM, Ray Rashif <schivmeister(a)gmail.com> wrote:
ah Ray, you and I have talked about this before and I completely
forgot. thanks for reminding me, I'll try using a different driver
and see if that helps (vesa, maybe?).
--
Josh Lawrence
http://www.hardbop200.com
23 Dec
23 Dec
3 a.m.
2009/12/22 Josh Lawrence <hardbop200(a)gmail.com>om>:
On Tue, Dec 22, 2009 at 12:14 AM, Ray Rashif
<schivmeister(a)gmail.com> wrote:
ah Ray, you and I have talked about this before and I completely
forgot. thanks for reminding me, I'll try using a different driver
and see if that helps (vesa, maybe?).
Ahh lol ok. But it'd really help if more people could comment on the
bug - that way it'll receive more attention and possibly lead to a
solution, better workaround, if not a patch.
VESA..hmm...sounds very useless =p
Now that the intel driver is so good, this had to happen :(
--
GPG/PGP ID: B42DDCAD
22 Dec
22 Dec
6:28 a.m.
2009/12/22 Josh Lawrence <hardbop200(a)gmail.com>om>:
hey list,
I've been struggling for the past couple of weeks trying to get my
netbook to perform well as a lightweight music workstation (mostly
synth work, light daw stuff). I'm getting many, many xruns, clients
getting kicked, etc. something is clearly not right. how do I begin
tracking down the source of xruns? I have an -rt kernel, a usb sound
card using -p256 -n3, running fluxbox with only wicd running to manage
my network. point is, I've tried the hunt-and-peck method and
(surprise) it hasn't done any good. how can I go about finding out
what exactly is tripping jack up, and what I can do to stop it?
Is your user a member of the audio group with rt permissions?
Have you tried pushing jackd's priority up? e.g. -P 70, on some
systems I've seemed to need that.
(PS - I've said it before: maybe I'm
expecting too much of a netbook.
could that be it?)
No, lots of people have done good things on netbooks. You don't say
which model you're on though.
Dan
3:48 p.m.
Dan S wrote:
2009/12/22 Josh Lawrence
<hardbop200(a)gmail.com>om>:
I believe we have a list member regularly using his EEEPC quite well in
performance, maybe he has some tips? I also seem to recall he had a link
where he put up his customized RT kernel packages optimized for EEEPC?
--
David
gnome(a)hawaii.rr.com
authenticity, honesty, community
hey list,
I've been struggling for the past couple of weeks trying to get my
netbook to perform well as a lightweight music workstation (mostly
synth work, light daw stuff). I'm getting many, many xruns, clients
getting kicked, etc. something is clearly not right. how do I begin
tracking down the source of xruns? I have an -rt kernel, a usb sound
card using -p256 -n3, running fluxbox with only wicd running to manage
my network. point is, I've tried the hunt-and-peck method and
(surprise) it hasn't done any good. how can I go about finding out
what exactly is tripping jack up, and what I can do to stop it?
Is your user a member of the audio group with rt permissions?
Have you tried pushing jackd's priority up? e.g. -P 70, on some
systems I've seemed to need that.
(PS - I've said it before: maybe I'm
expecting too much of a netbook.
could that be it?)
No, lots of people have done good things on netbooks. You don't say
which model you're on though.
4:06 p.m.
On Tue, Dec 22, 2009 at 1:48 PM, david <gnome(a)hawaii.rr.com> wrote:
I believe we have a list member regularly using his
EEEPC quite well in
performance, maybe he has some tips? I also seem to recall he had a link
where he put up his customized RT kernel packages optimized for EEEPC?
oh that's Ken Restivo (the best-ivo). I whine to him regularly off-list. :)
--
Josh Lawrence
http://www.hardbop200.com
23 Dec
23 Dec
3:36 a.m.
Josh Lawrence wrote:
On Tue, Dec 22, 2009 at 1:48 PM, david
<gnome(a)hawaii.rr.com> wrote:
And someday will there be a "Best-ivo of Ken Restivo" album out? ;-)
--
David
gnome(a)hawaii.rr.com
authenticity, honesty, community
I believe we have a list member regularly using
his EEEPC quite well in
performance, maybe he has some tips? I also seem to recall he had a link
where he put up his customized RT kernel packages optimized for EEEPC?
oh that's Ken Restivo (the best-ivo). I whine to him regularly off-list. :) 
26 Dec
26 Dec
2:35 p.m.
On Tue, Dec 22, 2009 at 09:35:58PM -1000, david wrote:
Josh Lawrence wrote:
Hah, as if anyone'd actually want to listen to such a thing!
The kernel and optimized packages I use to make my EEE go with Debian Lenny, are here:
http://restivo.org/projects/eee/debs/
Info on it is here:
http://restivo.org/projects/eee/
-ken
On Tue, Dec 22, 2009 at 1:48 PM, david
<gnome(a)hawaii.rr.com> wrote:
And someday will there be a "Best-ivo of Ken Restivo" album out? ;-)
I believe we have a list member regularly using
his EEEPC quite well in
performance, maybe he has some tips? I also seem to recall he had a link
where he put up his customized RT kernel packages optimized for EEEPC?
oh that's Ken Restivo (the best-ivo). I whine to him regularly off-list. :)
5488
days inactive
5493
days old
linux-audio-user@lists.linuxaudio.org
14 comments
7 participants
participants (7)
-
Dan S -
david -
Josh Lawrence -
Ken Restivo -
Ng Oon-Ee -
Patrick Shirkey -
Ray Rashif