13 Sep
2005
13 Sep
'05
12:29 p.m.
Hello,
I'm getting clicks in the audio using jack when there is video activity
(i.e. audio software that displays meters like Rezound, when muse
refreshes the display at playing and following the cursor, when
minimizing and maximizing windows, etc). If I increase the latency they
seem to stop. What could be causing this? Here there are some details of
my system:
Asus P4P800Deluxe
Pentium 4HT 2.8 ghz
1 gig ram
M-Audio Audiophile
Agnula Linux with kernel 2.6.12 (with preemption)
This is my cat /proc/interrupts
hector@demudi:~$ cat /proc/interrupts
CPU0
0: 3167206 IO-APIC-edge timer 0/67206
1: 601 IO-APIC-edge i8042 0/601
7: 0 IO-APIC-edge parport0 0/0
9: 0 IO-APIC-level acpi 0/0
14: 20770 IO-APIC-edge ide0 0/20769
15: 43 IO-APIC-edge ide1 0/41
177: 187389 IO-APIC-level uhci_hcd:usb1, radeon@pci:0000:01:00.0
0/87389185: 205449 IO-APIC-level uhci_hcd:usb2 0/5449
193: 3 IO-APIC-level ehci_hcd:usb3 0/3
201: 21466 IO-APIC-level Intel ICH5 0/21466
209: 3 IO-APIC-level ohci1394 0/3
217: 35511 IO-APIC-level SysKonnect SK-98xx 0/35511
225: 1696593 IO-APIC-level ICE1712 0/96593
NMI: 0
LOC: 3167251
ERR: 0
MIS: 0
I have jack set to 128 frames, 2 periods and 44100 sample rate.
The video card is a Radeon 7000.
As you can see my audio card is in the last IRQ (225: 1696593
IO-APIC-level ICE1712 0/96593) so I was wondering if this is the
cause. I tried to set a higher priority using chrt -f -p 99 `pidof "IRQ
225"` but it didn't fix anything.
I remember having the same problem long time ago in windows when I had
an Intel motherboard and I was using the built-in video card. The
problem disappeared when I installed a new AGP card.
Thank you!
Hector
13 Sep
13 Sep
3:06 p.m.
On Tue, 2005-09-13 at 10:29 -0400, Hector Centeno wrote:
Hello,
I'm getting clicks in the audio using jack when there is video activity
(i.e. audio software that displays meters like Rezound, when muse
refreshes the display at playing and following the cursor, when
minimizing and maximizing windows, etc). If I increase the latency they
seem to stop. What could be causing this? Here there are some details of
my system:
Asus P4P800Deluxe
Pentium 4HT 2.8 ghz
1 gig ram
M-Audio Audiophile
Agnula Linux with kernel 2.6.12 (with preemption)
Does JACK report xruns when the clicks happen?
Assuming this is a realtime-preempt kernel, what do you get
in /proc/latency_trace when you recompile the kernel with all latency
tracing options enabled?
Lee
14 Sep
14 Sep
12:49 a.m.
Lee Revell wrote:
On Tue, 2005-09-13 at 10:29 -0400, Hector Centeno
wrote:
Thank you for your reply. With Rezound I get maybe a couple of xruns
(and much more clicking) but in general the clicking doesn't cause any
xrun reports from jack (using other audio software, i.e. muse, ardour
and csound give no xruns at all but I get the clicking simultaneous with
graphic activity).
I'll try recompiling the kernel with the latency tracing options and
I'll let you know the results.
cheers,
Hector
Hello,
I'm getting clicks in the audio using jack when there is video activity
(i.e. audio software that displays meters like Rezound, when muse
refreshes the display at playing and following the cursor, when
minimizing and maximizing windows, etc). If I increase the latency they
seem to stop. What could be causing this? Here there are some details of
my system:
Asus P4P800Deluxe
Pentium 4HT 2.8 ghz
1 gig ram
M-Audio Audiophile
Agnula Linux with kernel 2.6.12 (with preemption)
Does JACK report xruns when the clicks happen?
Assuming this is a realtime-preempt kernel, what do you get
in /proc/latency_trace when you recompile the kernel with all latency
tracing options enabled?
Lee
4:12 a.m.
On Tue, 13 Sep 2005 22:49:13 -0400
Hector Centeno <h.centeno(a)sympatico.ca> wrote:
Thank you for your reply. With Rezound I get maybe a
couple of xruns
(and much more clicking) but in general the clicking doesn't cause any
xrun reports from jack (using other audio software, i.e. muse, ardour
and csound give no xruns at all but I get the clicking simultaneous with
graphic activity).
I'll try recompiling the kernel with the latency tracing options and
I'll let you know the results.
Hi,
are you using qjackctl?
If so, keep the connections window closed. the Jack graph gets locked
when its connections are queried, which qjackctl does quite often when
its connections window gets redrawn. This would cause 0-periods in the
audio stream but no xruns.
Regards,
Flo
--
Palimm Palimm!
http://tapas.affenbande.org
4:25 a.m.
Florian Schmidt wrote:
On Tue, 13 Sep 2005 22:49:13 -0400
Hector Centeno <h.centeno(a)sympatico.ca> wrote:
Yes, I'm using qjackctl and the problem is still present with the
connections window closed (I have the try icon activated, though).
Hector
Thank you for your reply. With Rezound I get maybe
a couple of xruns
(and much more clicking) but in general the clicking doesn't cause any
xrun reports from jack (using other audio software, i.e. muse, ardour
and csound give no xruns at all but I get the clicking simultaneous with
graphic activity).
I'll try recompiling the kernel with the latency tracing options and
I'll let you know the results.
Hi,
are you using qjackctl?
If so, keep the connections window closed. the Jack graph gets locked
when its connections are queried, which qjackctl does quite often when
its connections window gets redrawn. This would cause 0-periods in the
audio stream but no xruns.
Regards,
Flo
4:43 a.m.
On Wed, 14 Sep 2005 02:25:38 -0400
Hector Centeno <h.centeno(a)sympatico.ca> wrote:
>are you using qjackctl?
>
>If so, keep the connections window closed. the Jack graph gets locked
>when its connections are queried, which qjackctl does quite often when
>its connections window gets redrawn. This would cause 0-periods in the
>audio stream but no xruns.
Yes, I'm using qjackctl and the problem is still
present with the
connections window closed (I have the try icon activated, though).
Oh well, the tray icon shouldn't make any problems. It was just a wild
guess :)
Flo
--
Palimm Palimm!
http://tapas.affenbande.org
7:14 a.m.
Florian Schmidt wrote:
are you using qjackctl?
If so, keep the connections window closed. the Jack graph gets locked
when its connections are queried, which qjackctl does quite often when
its connections window gets redrawn. This would cause 0-periods in the
audio stream but no xruns.
FYI: qjackctl does NOT query jack connections whenever the window is
redrawn, so you may pretty well have the connections window open all the
time.
However, the qjackctl connections window get actually refreshed, by
querying and locking the jack graph instantaneously, only on these
particular moments:
1. when a new jack client enters or exits the graph;
2. when any active jack client changes its own our others
connections; this includes qjackctl normal but explicit user operation.
3. when the "Setup/Options/Auto refresh connections Patchbay, every
(secs) ..." option is active.
Although 1 and 2 are unavoidable, and as far as qjackctl is concerned it
should be better addressed by a lock-free jackd implementation (as
jackdmp promises), this later one (3) is not needed anymore and you
better leave it off.
Cheers.
--
rncbc aka Rui Nuno Capela
rncbc(a)rncbc.org
10:10 a.m.
On Wed, 14 Sep 2005 10:15:40 +0100
Rui Nuno Capela <rncbc(a)rncbc.org> wrote:
FYI: qjackctl does NOT query jack connections whenever
the window is
redrawn, so you may pretty well have the connections window open all the
time.
Ah, ok, my bad. I just had this impression once when i wondered where
the hell the audio dropouts (but no xruns) i experienced came from.
Closing the qjackctl connections window fixed it (i got a dropout
everytime i switched to the screen where qjackctl's connection window
was).
[snip]
Thanks for setting me straight and sorry for my spreading false infos.
Flo
--
Palimm Palimm!
http://tapas.affenbande.org
2:31 a.m.
Lee Revell wrote:
On Tue, 2005-09-13 at 10:29 -0400, Hector Centeno
wrote:
I was trying to build the kernel with the latency trace options and I
couldn't find CONFIG_LATENCY_TRACE in the config file. Does it have a
different name? I'm using the 2.6.12 source from the agnula repository.
Hector
Hello,
I'm getting clicks in the audio using jack when there is video activity
(i.e. audio software that displays meters like Rezound, when muse
refreshes the display at playing and following the cursor, when
minimizing and maximizing windows, etc). If I increase the latency they
seem to stop. What could be causing this? Here there are some details of
my system:
Asus P4P800Deluxe
Pentium 4HT 2.8 ghz
1 gig ram
M-Audio Audiophile
Agnula Linux with kernel 2.6.12 (with preemption)
Does JACK report xruns when the clicks happen?
Assuming this is a realtime-preempt kernel, what do you get
in /proc/latency_trace when you recompile the kernel with all latency
tracing options enabled?
Lee
7015
days inactive
7016
days old
linux-audio-user@lists.linuxaudio.org
8 comments
4 participants
participants (4)
-
Florian Schmidt -
Hector Centeno -
Lee Revell -
Rui Nuno Capela