15 Feb
2010
15 Feb
'10
12:18 p.m.
Hi,
I recently was testing out Seq24 >> Arpage >> Yoshimi, and it sounded
terribly out of time. This is on a fairly new Gentoo install so I
wondered if something was misconfigured. I also tried the first release
of Arpage and same result.
Then I noticed the latency of the Jack settings in use was quite high and
reset it to give a much lower latency which completely fixed the timing
problems the above seq/arp/synth setup was having.
The next test involved the Dino sequencer, but before doing so I
attempted to change the Jack settings to give a really bad latency,
170ms+. Jack could not start with such high latencies - the Alsa driver
would not load IIRC.
I reset the Jack settings for the worst latency I could get with Jack
starting, but the timing in Dino although far from perfect, was nowhere
near as bad as Arpage.
1) Is there any way to improve Jack Midi timing when the audio driver has
a high latency?
2) What is the reason the Alsa driver/Jack would not start when the
settings give a big 170ms+ latency?
BTW, I'm doing this stuff so I can program with Jack Midi.
TIA,
James.
15 Feb
15 Feb
12:20 p.m.
On Mon, Feb 15, 2010 at 11:18 AM, james morris <james(a)jwm-art.net> wrote:
1) Is there any way to improve Jack Midi timing when
the audio driver has
a high latency?
the timing has LOT to do with the ALSA MIDI "backend". If you use
a2jmidid, you will get dramatically better timing. i have already done
quite a lot of the work to make it possible to use a2jmidid "within"
JACK itself, but i'm not done yet. bet to get the git version of
2jmidid, since it has some important fixes.
2) What is the reason the Alsa driver/Jack would not
start when the
settings give a big 170ms+ latency?
probably your audio device does not support such a large buffer
--p
12:39 p.m.
On Mon, Feb 15, 2010 at 04:18:03PM +0000, james morris wrote:
1) Is there any way to improve Jack Midi timing when
the audio driver has
a high latency?
I don't think so. If you set the audio period size to
e.g. 4096, jack will call clients only every 85ms, and
midi gets delayed as well.
2) What is the reason the Alsa driver/Jack would not
start when the
settings give a big 170ms+ latency?
Probably because your soundcard doesn't have
such large buffers. Which is sort of reasonable.
BTW, I'm doing this stuff so I can program with
Jack Midi.
Nice, but why do you want such high latency ??
Ciao,
--
FA
O tu, che porte, correndo si ?
E guerra e morte !
1:07 p.m.
On 15/2/2010, "fons(a)kokkinizita.net" <fons(a)kokkinizita.net> wrote:
On Mon, Feb 15, 2010 at 04:18:03PM +0000, james morris
wrote:
I don't, it was just for testing out the hypothesis I thought would
explain why I was getting bad timing. I have just been testing it again,
and I actually think the problem was in Yoshimi.
Sorry for making Jack Midi sound bad :|
Running dino and yoshimi (both jack midi apps) the timing was bad even
with decent low latency (and then yoshimi zombified) but using phasex -
not a jack midi app (is it?) - (still with dino) the timing was near
perfect.
And thanks Paul for your answers too.
Cheers,
James.
1) Is there any way to improve Jack Midi timing
when the audio driver has
a high latency?
I don't think so. If you set the audio period size to
e.g. 4096, jack will call clients only every 85ms, and
midi gets delayed as well.
2) What is the reason the Alsa driver/Jack would
not start when the
settings give a big 170ms+ latency?
Probably because your soundcard doesn't have
such large buffers. Which is sort of reasonable.
BTW, I'm doing this stuff so I can program
with Jack Midi.
Nice, but why do you want such high latency ?? Ciao,
--
FA
O tu, che porte, correndo si ?
E guerra e morte !
_______________________________________________
Linux-audio-dev mailing list
Linux-audio-dev(a)lists.linuxaudio.org
http://lists.linuxaudio.org/listinfo/linux-audio-dev
1:31 p.m.
On Mon, Feb 15, 2010 at 11:35 AM, <fons(a)kokkinizita.net> wrote:
On Mon, Feb 15, 2010 at 04:18:03PM +0000, james morris
wrote:
fons, recall the difference between latency and jitter. the problem
with the -Xseq backend (and -Xraw too) is that it causes really bad
jitter in addition to whatever latency is present. a2jmidid (at least
the current git version) doesn't do this - it keeps the jitter as low
as possible and the latency is constant.
1) Is there any way to improve Jack Midi timing
when the audio driver has
a high latency?
I don't think so. If you set the audio period size to
e.g. 4096, jack will call clients only every 85ms, and
midi gets delayed as well.
2:36 p.m.
On Mon, Feb 15, 2010 at 12:31:19PM -0500, Paul Davis wrote:
On Mon, Feb 15, 2010 at 11:35 AM,
<fons(a)kokkinizita.net> wrote:
Yes, you're right. Since midi events get timestamped w.r.t.
the start of the current period a large period just adds
latency and not jitter.
So the solution would be a separate thread that receives
MIDI all the time and not just at the start of a cycle ?
Ciao,
--
FA
O tu, che porte, correndo si ?
E guerra e morte !
On Mon, Feb 15, 2010 at 04:18:03PM +0000, james
morris wrote:
fons, recall the difference between latency and jitter. the problem
with the -Xseq backend (and -Xraw too) is that it causes really bad
jitter in addition to whatever latency is present. a2jmidid (at least
the current git version) doesn't do this - it keeps the jitter as low
as possible and the latency is constant. 1) Is there any way to improve Jack Midi timing
when the audio driver has
a high latency?
I don't think so. If you set the audio period size to
e.g. 4096, jack will call clients only every 85ms, and
midi gets delayed as well.
2:40 p.m.
On Mon, Feb 15, 2010 at 1:31 PM, <fons(a)kokkinizita.net> wrote:
So the solution would be a separate thread that
receives
MIDI all the time and not just at the start of a cycle ?
a2jmidid already implements this, and i have working internal JACK
clients based on it that do the same.
--p
4:44 p.m.
On Mon, Feb 15, 2010 at 1:31 PM,
<fons(a)kokkinizita.net> wrote:
Is this MIDI thread just internal to Jack or is it the audio client that needs the
extra thread?
Kind regards, nick
"we have to make sure the old choice [Windows] doesn't disappear”.
Jim Wong, president of IT products, Acer
_________________________________________________________________
Your E-mail and More On-the-Go. Get Windows Live Hotmail Free.
https://signup.live.com/signup.aspx?id=60969
So the solution would be a separate thread that
receives
MIDI all the time and not just at the start of a cycle ?
<paul(a)linuxaudiosystems.com> wrote:
a2jmidid already implements this, and i have working internal JACK
clients based on it that do the same.
4:49 p.m.
On Mon, Feb 15, 2010 at 3:44 PM, Nick Copeland
<nickycopeland(a)hotmail.com> wrote:
for a2jmidid, its just an extra thread in that client.
for my spinoff of this, its an extra thread in the server.
in case does it involve extra threads in normal JACK clients. JACK
MIDI data is passed around in JACK ports via process(), same as audio.
this merely concerns the points where MIDI data enters and leaves the
JACK graph for uncharted territory (e.g. ALSA sequencer, CoreMIDI, etc
etc)
On Mon, Feb
15, 2010 at 1:31 PM, <fons(a)kokkinizita.net> wrote:
Is this MIDI thread just internal to Jack or is it the audio client that
needs the
extra thread? So the solution would be a separate thread that
receives
MIDI all the time and not just at the start of a cycle ?
<paul(a)linuxaudiosystems.com> wrote:
a2jmidid already implements this, and i have working internal JACK
clients based on it that do the same.
5188
days inactive
5188
days old
linux-audio-dev@lists.linuxaudio.org
8 comments
4 participants
participants (4)
-
fons＠kokkinizita.net -
james morris -
Nick Copeland -
Paul Davis