2 May
2013
2 May
'13
10:48 p.m.
Hello everyone!
I've got a problem uploading some sounds to my Waldorf Microwave XT. It's an
old device, but it's made to receive patches and wavetables. I downloaded a
couple of MIDI fies containing the necessary SysEx data from the web. But when
using aplaymidi to upload them like this:
aplaymidi -p 24:0 bank.mid
I get nothing for a few seconds and then messages like this get repeated a
couple of times:
[1393510.800523] dump_midi: 20 callbacks suppressed
In the end I don't find any difference in my synth.
question: is it the synth, the MIDI interface or ALSA/aplaymidi. My MIDI
interface in this instance is a USB -> MIDI adapter, just shown as "USB MIDI
Interface". It's twin is called "USB MIDI Cable". So I can't tell
you it's
make and model. So far they've both served me well.
Any suggestions?
Kindest regards
Julien
----------------------------------------
http://juliencoder.de/nama/music.html
3 May
3 May
8:48 a.m.
Julien Claassen wrote:
dump_midi: 20 callbacks suppressed
The interesting message is the one before that.
My crystal ball says it's "MIDI output buffer overrun".
This happens when some program tries to send data at a higher rate than
MIDI allows, and when the additional buffer that ALSA uses to smooth
over such spikes overflows.
Your .mid file is horribly broken.
Change the .mid file to insert proper delays between the SysEx commands,
or use a .syx file and send it with amidi.
(To convert a .mid file to .syx, run "amidi -p virtual -r myfile.syx"
and play the .mid file with aplaymidi to the virtual port.)
Regards,
Clemens
9:25 a.m.
Hi Clemens!
Thanks for the tip. I'll try to convert the MIDI file to .syx.
A question: what is so special about amidi? I mean: what is the difference
between it playing a pure sysex file and aplaymidi playing the sysex block
from a MIDI file?
Is there a commandline tool to dump sysex like that, using the ALSA
sequencer system, by any chance?
Warm regards
Julien
----------------------------------------
http://juliencoder.de/nama/music.html
9:37 a.m.
When receiving a normal 3 byte MIDI message most gear can process it in the interval
between messages. When receiving a large sysex buffer, a lot of gear chokes on converting
the MSG byte by byte into something usable - either because of a low buffer size, or some
horrible code on the part of the sysex parser.
Slowing sysex messages down has nothing to do with the baud rate of MIDI but more to do
with the between message gap and the processing power available during interrupt handling
periods. Slowing a sysex dump down means "just pause a lot more between bytes so that
the destination CPU can keep up"..
;
--
seclorum
On May 3, 2013, at 9:25, Julien Claassen <julien(a)mail.upb.de> wrote:
Hi Clemens!
Thanks for the tip. I'll try to convert the MIDI file to .syx.
A question: what is so special about amidi? I mean: what is the difference between it
playing a pure sysex file and aplaymidi playing the sysex block from a MIDI file?
Is there a commandline tool to dump sysex like that, using the ALSA sequencer system, by
any chance?
Warm regards
Julien
----------------------------------------
http://juliencoder.de/nama/music.html
_______________________________________________
Linux-audio-user mailing list
Linux-audio-user(a)lists.linuxaudio.org
http://lists.linuxaudio.org/listinfo/linux-audio-user
10:22 a.m.
Julien Claassen wrote:
what is so special about amidi? I mean: what is the
difference between
it playing a pure sysex file and aplaymidi playing the sysex block
from a MIDI file?
.mid files contain timing information for each message; aplaymidi
schedules each message to be sent at the specified time. If the MIDI
port's buffer has not enough free space for the message, there is
nowhere else to store it, so it gets dropped.
amidi accesses the MIDI port as a dumb serial pipe. There is not timing
information in .syx files, so amidi sends the data as fast as possible,
but when the buffer is full, amidi will wait.
Is there a commandline tool to dump sysex like that,
using the ALSA
sequencer system, by any chance?
That "virtual" port you can create with amidi _is_ using the ALSA
sequencer system.
Regards,
Clemens
8:44 p.m.
On Fri, 3 May 2013, Clemens Ladisch wrote:
Julien Claassen wrote:
Previous posters are correct that it's easiest to send sysex files with
the amidi command, plus there's a nice Java program called JSynthLib
that will function as a generic sysex librarian plus it has some
specific editing capabilities for certain synthesizers:
http://sourceforge.net/p/jsynthlib/wiki/Home/
(JSynthLib is great if you have hundreds of banks worth of Ensoniq ESQ-1
patches...)
The problem I have with both amidi and JSynthLib though is that because
they both try to access the Alsa Midi device raw, they both get blocked
with "in use" messages when jackd is running, which has the effect that
I have to be sure to schedule my time so as to never do sysex dumps when
any Jack application is running (which is annoying).
Is there any way to do this kind of raw Alsa access with jackd going?
Note that this is not the same thing as programs like Hydrogen that
create Midi ports that show up on the Alsa tab in the QJackCtl
connection manager... Programs that do raw access like this don't show
up as they can't even launch -- they just get blocked with an error.
It'd be nice to be able to sysex any time you want to.
--
+ Brent A. Busby + "We've all heard that a million monkeys
+ Sr. UNIX Systems Admin + banging on a million typewriters will
+ University of Chicago + eventually reproduce the entire works of
+ James Franck Institute + Shakespeare. Now, thanks to the Internet,
+ Materials Research Ctr + we know this is not true." -Robert Wilensky
what is so special about amidi? I mean: what is
the difference
between it playing a pure sysex file and aplaymidi playing the sysex
block from a MIDI file?
.mid files contain timing information for each message; aplaymidi
schedules each message to be sent at the specified time. If the MIDI
port's buffer has not enough free space for the message, there is
nowhere else to store it, so it gets dropped.
amidi accesses the MIDI port as a dumb serial pipe. There is not
timing information in .syx files, so amidi sends the data as fast as
possible, but when the buffer is full, amidi will wait.
Is there a commandline tool to dump sysex like
that, using the ALSA
sequencer system, by any chance?
That "virtual" port you can create with amidi _is_ using the ALSA
sequencer system.
9:41 p.m.
Hello Brent!
I don't think so. There is another application, which is just as useless as
jsynthlib, but might be of help to you: simple sysexxer. Not sure, what it
uses to access the soundcard.
Warm regards
Julien
----------------------------------------
http://juliencoder.de/nama/music.html
4 May
4 May
12:41 a.m.
Hi,
On 3 May 2013 19:44, "Brent Busby" <brent(a)keycorner.org> wrote:
Is there any way to do this kind of raw Alsa access
with jackd going?
You might be able to use the vir-midi module for this. I looked into this
to help someone get JACK MIDI into Praxis. Should work for output too.
Bit more about what I did here -
http://code.google.com/p/praxis/issues/detail?id=31
Best wishes,
Neil
Neil C Smith
Artist : Technologist : Adviser
http://neilcsmith.net
5:06 p.m.
On Fri, 3 May 2013, Neil C Smith wrote:
On 3 May 2013 19:44, "Brent Busby"
<brent(a)keycorner.org> wrote:
I checked into this, and unless I'm misunderstanding what the module
does, it doesn't look like it will help. It seems to provide a virtual
port for OSS apps to be able to access the Alsa sequencer ports. The
problem is amidi is already a native Alsa program, but it can't access
the hardware ports because Jack has already grabbed them. It's not an
OSS versus Alsa issue, but just that the device is exclusively in use.
It almost seems like what we need is some kind of Jack-aware amidi
command...
--
+ Brent A. Busby + "We've all heard that a million monkeys
+ Sr. UNIX Systems Admin + banging on a million typewriters will
+ University of Chicago + eventually reproduce the entire works of
+ James Franck Institute + Shakespeare. Now, thanks to the Internet,
+ Materials Research Ctr + we know this is not true." -Robert Wilensky
Is there any way to do this kind of raw Alsa
access with jackd going?
You might be able to use the vir-midi module for this. I looked into this
to help someone get JACK MIDI into Praxis. Should work for output too.
Bit more about what I did here -
http://code.google.com/p/praxis/issues/detail?id=31 
5:20 p.m.
JACK does not "grab" anything except *raw* MIDI ports and even then only if
-Xraw is used, which is a bad idea to begin with.
On Sat, May 4, 2013 at 11:06 AM, Brent Busby <brent(a)keycorner.org> wrote:
On Fri, 3 May 2013, Neil C Smith wrote:
On 3 May 2013 19:44, "Brent Busby" <brent(a)keycorner.org> wrote:
I checked into this, and unless I'm misunderstanding what the module does,
it doesn't look like it will help. It seems to provide a virtual port for
OSS apps to be able to access the Alsa sequencer ports. The problem is
amidi is already a native Alsa program, but it can't access the hardware
ports because Jack has already grabbed them. It's not an OSS versus Alsa
issue, but just that the device is exclusively in use.
It almost seems like what we need is some kind of Jack-aware amidi
command...
--
+ Brent A. Busby + "We've all heard that a million monkeys
+ Sr. UNIX Systems Admin + banging on a million typewriters will
+ University of Chicago + eventually reproduce the entire works of
+ James Franck Institute + Shakespeare. Now, thanks to the Internet,
+ Materials Research Ctr + we know this is not true." -Robert Wilensky
______________________________**_________________
Linux-audio-user mailing list
Linux-audio-user@lists.**linuxaudio.org<Linux-audio-user@lists.linuxaudio.org>
http://lists.linuxaudio.org/**listinfo/linux-audio-user<http://lists.lin…
Is there any way to do this kind of raw Alsa
access with jackd going?
You might be able to use the vir-midi module for this. I looked into this
to help someone get JACK MIDI into Praxis. Should work for output too.
Bit more about what I did here -
http://code.google.com/p/**praxis/issues/detail?id=31<http://code.google…
5:39 p.m.
On Sat, 4 May 2013, Paul Davis wrote:
JACK does not "grab" anything except *raw*
MIDI ports and even then
only if -Xraw is used, which is a bad idea to begin with.
I remember reading that before, but the strange thing is I'm setup for
"seq" in QJackCtl, and my ~/.jackdrc has "-Xseq", and yet if I try to
send sysex from the shell with amidi while jackd is running I get a
device in use error. I don't know how to explain it, but that's what it
does, even with -Xseq.
--
+ Brent A. Busby + "We've all heard that a million monkeys
+ Sr. UNIX Systems Admin + banging on a million typewriters will
+ University of Chicago + eventually reproduce the entire works of
+ James Franck Institute + Shakespeare. Now, thanks to the Internet,
+ Materials Research Ctr + we know this is not true." -Robert Wilensky
5 May
5 May
12:58 a.m.
I'd like to get the snd-virmidi driver working on my setup, but even
after reading the documentation included with the kernel for Alsa and
the Alsa wiki, I'm a bit murky on all the stuff that's in modprobe's
configuration for getting everything loaded at bootup. Currently, I
have an Intel sound chipset on my motherboard that I'm using index=0 to
ensure will be my primary sound device, followed by my RME Multiface
with index=1. I also have an Edirol UM-3EX USB 3-port Midi interface
that's being forced to be last with index=-2, since it's not really even
a sound device and I want it to be last (though how "-2" is achieving
this is one of the things I'm murky about, and hesitant to screw with).
This yields the following presently positive result in
/proc/asound/cards:
village:~# cat /proc/asound/cards
0 [SB ]: HDA-Intel - HDA ATI SB
HDA ATI SB at 0xfb9f8000 irq 16
1 [DSP ]: H-DSP - Hammerfall DSP
RME Hammerfall DSP + Multiface at 0xfbbf0000, irq 19
2 [UM3EX ]: USB-Audio - UM-3EX
EDIROL UM-3EX at usb-0000:00:12.1-2.1, full speed
If I enable the snd-virmidi driver, I'd like it to be even lower in
priority than the Edirol UM-3EX Midi, and be last. Probably that means
the Edirol now becomes index=2, but does snd-virmidi become index=-2?
The kernel documentation says negative card indexes are "a bitmask of
permissable indexes"...from what? An 8-bit number? How does that work
exactly? I do understand what a bitmask is...but what is this doing
really? With my current configuration, I'm getting what I want, but I'm
unsettled that I don't know how it's being achieved. Just for
amusement/advice, I've included my /etc/modprobe.d/alsa.conf below.
Feel free to point out anything that looks foolish, because it's largely
been adapted from what the modprobe package came with by default:
-------------------------------------------------------------------------
# Alsa kernel modules' configuration file.
# ALSA portion
alias char-major-116 snd
# OSS/Free portion
alias char-major-14 soundcore
##
## IMPORTANT:
## You need to customise this section for your specific sound card(s)
## and then run `update-modules' command.
## Read alsa-driver's INSTALL file in /usr/share/doc for more info.
##
## ALSA portion
alias snd-card-1 snd-hdsp
alias snd-card-0 snd-hda-intel
options snd-hda-intel index=0
options snd-hdsp index=1
options snd-usb-audio index=-2
## OSS/Free portion
alias sound-slot-0 snd-card-0
alias sound-slot-1 snd-card-1
##
# OSS/Free portion - card #1
alias sound-service-0-0 snd-mixer-oss
alias sound-service-0-1 snd-seq-oss
alias sound-service-0-3 snd-pcm-oss
alias sound-service-0-8 snd-seq-oss
alias sound-service-0-12 snd-pcm-oss
# OSS/Free portion - card #2
alias sound-service-1-0 snd-mixer-oss
alias sound-service-1-3 snd-pcm-oss
alias sound-service-1-12 snd-pcm-oss
alias /dev/mixer snd-mixer-oss
alias /dev/dsp snd-pcm-oss
alias /dev/midi snd-seq-oss
# Set this to the correct number of cards.
options snd cards_limit=4
# options snd slots=snd-hda-intel,snd-hdsp,snd-usb-audio
--
+ Brent A. Busby + "We've all heard that a million monkeys
+ Sr. UNIX Systems Admin + banging on a million typewriters will
+ University of Chicago + eventually reproduce the entire works of
+ James Franck Institute + Shakespeare. Now, thanks to the Internet,
+ Materials Research Ctr + we know this is not true." -Robert Wilensky
12:42 p.m.
New subject: sound card order (was: Problem uploading sysex data to hardware synth)
Brent Busby wrote:
I also have an Edirol UM-3EX USB 3-port Midi interface
that's being
forced to be last with index=-2, since it's not really even a sound
device and I want it to be last (though how "-2" is achieving this is
one of the things I'm murky about, and hesitant to screw with).
As a 32-bit two's-complement negative integer, -2 in binary is
11111111111111111111111111111110. This means that the snd-usb-audio
driver is allowed to attach to any index except 0. (If the USB driver
is loaded before the HDSP driver, you lose.)
If I enable the snd-virmidi driver, I'd like it to
be even lower in
priority than the Edirol UM-3EX Midi, and be last.
Do not use the index= option. Always use the slots= options of the snd
module instead.
options snd-... index=...
...
# options snd slots=snd-hda-intel,snd-hdsp,snd-usb-audio
Remove the index= lines, uncomment the last line, and add the
snd-virmidi driver where desired.
Regards,
Clemens
3:17 p.m.
New subject: sound card order (was: Problem uploading sysex data to hardware synth)
On Sun, 5 May 2013, Clemens Ladisch wrote:
Brent Busby wrote:
Thank you very much for explaining that... It looks like I must have
been looking into that at some point, but commented it out instead of
trying it. It's certainly looks a lot easier with slots=.
--
+ Brent A. Busby + "We've all heard that a million monkeys
+ Sr. UNIX Systems Admin + banging on a million typewriters will
+ University of Chicago + eventually reproduce the entire works of
+ James Franck Institute + Shakespeare. Now, thanks to the Internet,
+ Materials Research Ctr + we know this is not true." -Robert Wilensky
I also have an Edirol UM-3EX USB 3-port Midi
interface that's being
forced to be last with index=-2, since it's not really even a sound
device and I want it to be last (though how "-2" is achieving this is
one of the things I'm murky about, and hesitant to screw with).
As a 32-bit two's-complement negative integer, -2 in binary is
11111111111111111111111111111110. This means that the snd-usb-audio
driver is allowed to attach to any index except 0. (If the USB driver
is loaded before the HDSP driver, you lose.)
If I enable the snd-virmidi driver, I'd like
it to be even lower in
priority than the Edirol UM-3EX Midi, and be last.
Do not use the index= option. Always use the slots= options of the snd
module instead.
options snd-... index=...
...
# options snd slots=snd-hda-intel,snd-hdsp,snd-usb-audio
Remove the index= lines, uncomment the last line, and add the
snd-virmidi driver where desired.
4 May
4 May
6:04 p.m.
Hi,
On 4 May 2013 16:06, Brent Busby <brent(a)keycorner.org> wrote:
On Fri, 3 May 2013, Neil C Smith wrote:
I was replying more in the context of JSynthLib / JavaSound as Praxis
LIVE's MIDI support is Java based. Should be the same with amidi
though if it only works with raw MIDI devices. vir-midi creates
virtual hardware devices that can be selected as outputs from programs
that only support raw MIDI, and makes them appear as ALSA devices that
you can connect via QJackCtl, etc.. If you've read the linked bug
report, you'll also notice I suggest to use a2jmidid which then makes
the virtual hardware devices appear as JACK MIDI devices.
You can try setting JSynthLib or amidi to play to one of the virtual
hardware outputs. IIRC this then loops back to the input device which
you can connect to whatever you want.
Hope that makes sense.
Best wishes,
Neil
--
Neil C Smith
Artist : Technologist : Adviser
http://neilcsmith.net
Praxis LIVE - open-source, graphical environment for rapid development
of intermedia performance tools, projections and interactive spaces -
http://code.google.com/p/praxis
OpenEye - specialist web solutions for the cultural, education,
charitable and local government sectors - http://openeye.info
You might be able to use the vir-midi module for
this. I looked into this
to help someone get JACK MIDI into Praxis. Should work for output too.
Bit more about what I did here -
http://code.google.com/p/praxis/issues/detail?id=31
I checked into this, and unless I'm misunderstanding what the module does,
it doesn't look like it will help. It seems to provide a virtual port for
OSS apps to be able to access the Alsa sequencer ports. The problem is
amidi is already a native Alsa program, but it can't access the hardware
ports because Jack has already grabbed them. It's not an OSS versus Alsa
issue, but just that the device is exclusively in use.
5 May
5 May
11:43 a.m.
On Fri, 2013-05-03 at 13:44 -0500, Brent Busby wrote:
Thank you for the information, very interesting for me since it does
allows editing of the Matrix 1000 and MT-32 :) + managing DX7
libraries :).
Perhaps not useful for Julien, since (as Simple Sysexxer too) it seems
to be GUI based.
Regards,
Ralf
3 May
3 May
8:49 a.m.
Quoting Julien Claassen <julien(a)mail.upb.de>de>:
I downloaded a couple of MIDI fies containing the
necessary SysEx data
from the web. But when using aplaymidi to upload them like this:
aplaymidi -p 24:0 bank.mid
I get nothing for a few seconds
Use amidi instead. It's designed to send and recieve sysex and other
"raw" MIDI data.
- Peder
9:27 a.m.
Hello Peder!
Thanks for the hint, but amidi will only play pure sysex files. I tried to
play a MIDI file and got the message, that this file should be played using
aplaymidi.
Best wishes
Julien
----------------------------------------
http://juliencoder.de/nama/music.html
1:40 p.m.
On Fri, May 03, 2013 at 09:27:53AM +0200, Julien Claassen wrote:
Hello Peder!
Thanks for the hint, but amidi will only play pure sysex files. I
tried to play a MIDI file and got the message, that this file should
be played using aplaymidi.
Will midish handle this?
Best wishes
Julien
--
Joel Roth
1:47 p.m.
Hi Joel1
Midish had the same problem as aplaymidi, even though I tried taking the
tempo down. But I managed something with amidi now.
Again: thanks Clemens for the conversion tip.
Unfortunately, there are still a few problems, but these seem to be
connected to the instrument specifics. at least I managed something.
Warm regards
Julien
----------------------------------------
http://juliencoder.de/nama/music.html
4294
days inactive
4297
days old
linux-audio-user@lists.linuxaudio.org
21 comments
9 participants
participants (9)
-
Brent Busby -
Clemens Ladisch -
Jay Vaughan -
Joel Roth -
Julien Claassen -
Neil C Smith -
Paul Davis -
Peder Hedlund -
Ralf Mardorf