17 Jan
2006
17 Jan
'06
10:26 a.m.
Greetings:
I've had nothing but bad luck with the snd-virmidi module and the
Demudi 2.6 kernels. Alas, no-one seems to be able to say what the
problem might be, so I thought I'd ask here: Does anyone have
snd-virmidi actually working on any Demudi 2.6 kernel ? On my system
(1.3.0rc1) it will load correctly, it shows up in QJackCtl's MIDI
connections panel, but it simply doesn't work. No MIDI is passing
through any virmidi port. MIDI does work perfectly via the MIDI Through
and via the hardware port, so I know it's not a general system error.
Btw, if you have loaded it please be sure to actually test it.
And just to clarify: MIDI Through will *not* do the job I need done. I
need the virmidi ports.
This is the second Demudi kernel that has had serious problems with
this module. I would like to hear from anyone who has it working under
Demudi. I'd really like to hear from anyone who has a clue as to why
it's so fubar'd here.
Best,
dp
17 Jan
17 Jan
1:29 p.m.
Dave Phillips wrote:
I've had nothing but bad luck with the
snd-virmidi module and the
Demudi 2.6 kernels.
Did you have any kernel version that worked?
On my system it will load correctly, it shows up in
QJackCtl's MIDI
connections panel, but it simply doesn't work. No MIDI is passing
through any virmidi port.
Does it work with aseqdump+amidi? If not, please show the exact commands
you used to test it.
Regards,
Clemens
1:47 p.m.
Clemens Ladisch wrote:
I tested it with seq24, hooking its output to the virmidi ports from
within seq24 itself (that's how it works). The virmidi port(s) were
connected to qsynth and to the hardware synth on the SBLive. I started
seq24, got no sound. I hooked the output to the MIDI Through port, got
sound. I also performed the same tests with the same hookups, but using
a MIDI keyboard to send events to the virmidi ports. Got the same results.
I tested it just now with aseqdump, again with the same results. MIDI
Through works fine, the virmidi ports report connection but no i/o.
Best,
dp
Dave Phillips wrote:
Sure, it's flawless on the 2.4 kernels.
I've had nothing but bad luck with the
snd-virmidi module and the
Demudi 2.6 kernels.
Did you have any kernel version that worked?
On my system it
will load correctly, it shows up in QJackCtl's MIDI
connections panel, but it simply doesn't work. No MIDI is passing
through any virmidi port.
Does it work with aseqdump+amidi? If not, please show the exact commands
you used to test it.
2:14 p.m.
Dave Phillips wrote:
Clemens Ladisch wrote:
You never mentioned the raw MIDI port (which es where the MIDI data shows up).
What do you want to do that the through ports cannot do?
Regards,
Clemens
Does it work with aseqdump+amidi? If not, please
show the exact commands
you used to test it.
I tested it with seq24, hooking its output to the virmidi ports from
within seq24 itself (that's how it works). The virmidi port(s) were
connected to qsynth and to the hardware synth on the SBLive. I started
seq24, got no sound. I hooked the output to the MIDI Through port, got
sound. I also performed the same tests with the same hookups, but using
a MIDI keyboard to send events to the virmidi ports. Got the same results.
2:54 p.m.
Clemens Ladisch wrote:
You never mentioned the raw MIDI port (which es where the MIDI data shows up).
I'm not sure what you mean here. I've attached a variety of reports to
the end of this message, perhaps you'll see something significant there.
Does it work with aseqdump+amidi? If not, please show
the exact commands
you used to test it.
I tested it with seq24, hooking its output to the virmidi ports from
within seq24 itself (that's how it works). The virmidi port(s) were
connected to qsynth and to the hardware synth on the SBLive. I started
seq24, got no sound. I hooked the output to the MIDI Through port, got
sound. I also performed the same tests with the same hookups, but using
a MIDI keyboard to send events to the virmidi ports. Got the same results.
What do you want to do that the through ports cannot
do?
I use Xsteem and Dosemu to run various MIDI applications. They want a
direct link to the device, and linking to /dev/snd/midiCxDx gives me the
greatest flexibility. It would be much better if they abided by the ALSA
convention of indicating devices (hw:x,x), but they don't.
Best,
dp
******************** Reports from Demudi 1.3.0rc1 **************
dlphilp@localdomain:~$ cat /proc/asound/cards
0 [M66 ]: ICE1712 - M Audio Delta 66
M Audio Delta 66 at 0xd800, irq 5
1 [Live ]: EMU10K1 - SBLive! Value [CT4832]
SBLive! Value [CT4832] (rev.8, serial:0x80271102)
at 0xd000, irq 3
2 [VirMIDI ]: VirMIDI - VirMIDI
Virtual MIDI Card 1
dlphilp@localdomain:~$ lsmod |grep midi
snd_virmidi 4288 1
snd_seq_virmidi 8448 3 snd_virmidi,snd_emux_synth
snd_seq_midi_emul 7680 1 snd_emux_synth
snd_seq_midi 9504 2
snd_rawmidi 26528 4
snd_seq_virmidi,snd_emu10k1,snd_mpu401_uart,snd_seq_midi
snd_seq_midi_event 7680 3 snd_seq_virmidi,snd_seq_oss,snd_seq_midi
snd_seq 56336 13
snd_emux_synth,snd_seq_virmidi,snd_seq_midi_emul,snd_seq_instr,snd_seq_dummy,snd_seq_oss,snd_seq_midi,snd_seq_midi_event
snd_seq_device 8844 8
snd_emu10k1_synth,snd_emux_synth,snd_emu10k1,snd_seq_dummy,snd_seq_oss,snd_seq_midi,snd_rawmidi,snd_seq
snd 58596 24
snd_virmidi,snd_emux_synth,snd_seq_virmidi,snd_emu10k1,snd_hwdep,snd_ice1712,snd_ak4xxx_adda,snd_cs8427,snd_ac97_codec,snd_pcm_oss,snd_mixer_oss,snd_pcm,snd_i2c,snd_mpu401_uart,snd_seq_oss,snd_rawmidi,snd_seq,snd_timer,snd_seq_device
dlphilp@localdomain:~$ amidi -l
Device Name
hw:1,0 EMU10K1 MPU-401 (UART)
hw:1,1 Emu10k1 Synth MIDI (16 subdevices)
hw:1,2 Emu10k1 Synth MIDI (16 subdevices)
hw:2,0 Virtual Raw MIDI (16 subdevices)
hw:2,1 Virtual Raw MIDI (16 subdevices)
hw:2,2 Virtual Raw MIDI (16 subdevices)
hw:2,3 Virtual Raw MIDI (16 subdevices)
dlphilp@localdomain:~$ aseqdump -l
Port Client name Port name
0:0 System Timer
0:1 System Announce
62:0 Midi Through Midi Through Port-0
72:0 EMU10K1 MPU-401 (UART) EMU10K1 MPU-401 (UART)
80:0 Virtual Raw MIDI 2-0 VirMIDI 2-0
81:0 Virtual Raw MIDI 2-1 VirMIDI 2-1
82:0 Virtual Raw MIDI 2-2 VirMIDI 2-2
83:0 Virtual Raw MIDI 2-3 VirMIDI 2-3
hw {
@args.0 CARD
@args.1 DEV
@args.2 SUBDEV
@args.CARD {
type string
default {
@func getenv
vars {
0 ALSA_RAWMIDI_CARD
1 ALSA_CARD
}
default {
@func refer
name 'defaults.rawmidi.card'
}
}
}
@args.DEV {
type integer
default {
@func igetenv
vars {
0 ALSA_RAWMIDI_DEVICE
}
default {
@func refer
name 'defaults.rawmidi.device'
}
}
}
@args.SUBDEV {
type integer
default -1
}
type hw
card $CARD
device $DEV
subdevice $SUBDEV
}
default {
type hw
card {
@func getenv
vars {
0 ALSA_RAWMIDI_CARD
1 ALSA_CARD
}
default {
@func refer
name 'defaults.rawmidi.card'
}
}
device {
@func igetenv
vars {
0 ALSA_RAWMIDI_DEVICE
}
default {
@func refer
name 'defaults.rawmidi.device'
}
}
}
virtual {
@args.0 MERGE
@args.MERGE {
type string
default 1
}
type virtual
merge $MERGE
}
dlphilp@localdomain:~$ modinfo snd-virmidi
filename:
/lib/modules/2.6.14-1-multimedia-386/kernel/sound/drivers/snd-virmidi.ko
author: Takashi Iwai <tiwai(a)suse.de>
description: Dummy soundcard for virtual rawmidi devices
license: GPL
vermagic: 2.6.14-1-multimedia-386 preempt 686 gcc-4.0
depends: snd,snd-seq-virmidi
parm: midi_devs:MIDI devices # (1-8) (array of int)
parm: enable:Enable this soundcard. (array of bool)
parm: id:ID string for virmidi soundcard. (array of charp)
parm: index:Index value for virmidi soundcard. (array of int)
18 Jan
18 Jan
4:45 a.m.
Dave Phillips wrote:
Clemens Ladisch wrote:
The raw MIDI ports are the old OSS device files (/dev/midi*).
In your configuration, the virtual MIDI devices are configured like this:
OSS emul. ALSA raw MIDI ALSA sequencer
/dev/midi02 hw:2,0 <--> 80:0
/dev/amidi02 hw:2,1 <--> 80:1
hw:2,2 <--> 80:2
hw:2,3 <--> 80:3
Please run the following commands (in separate consoles):
aseqdump -p 80:0
amidi -p hw:2,0 -S "90 12 34 80 56 78"
or, for the other direction:
amidi -p hw:2,0 -d
aplaymidi -p 80:0 something.mid
Is there any output?
Regards,
Clemens
You never mentioned the raw MIDI port (which es
where the MIDI data
shows up).
I'm not sure what you mean here.
9:28 a.m.
Clemens Ladisch wrote:
The raw MIDI ports are the old OSS device files (/dev/midi*).
In your configuration, the virtual MIDI devices are configured like this:
OSS emul. ALSA raw MIDI ALSA sequencer
/dev/midi02 hw:2,0 <--> 80:0
/dev/amidi02 hw:2,1 <--> 80:1
hw:2,2 <--> 80:2
hw:2,3 <--> 80:3
Thank you for that clarification. So the ALSA virtual MIDI ports are
mapped to the raw MIDI ports ?
You never mentioned the raw MIDI port (which es where
the MIDI data
shows up).
I'm not sure what you mean here.
Please run the following commands (in separate
consoles):
aseqdump -p 80:0
amidi -p hw:2,0 -S "90 12 34 80 56 78"
or, for the other direction:
amidi -p hw:2,0 -d
aplaymidi -p 80:0 something.mid
Is there any output?
Yes, there is. But if I connect the virmidi port to a synth there's
still no sound. The second example shows the MIDI file playing along
fine, but using QJackCtl to connect 80:0 to a synth produces no sound. I
set the port to the MIDI Thru port, got sound. How can this be happening ?
Btw, when I rebooted into Demudi this morning the virmidi card was
placed first in my cards list in /proc/asound/cards. I removed the
module and tried to reload it with index=2 but the load failed with an
error stating that snd-virmidi.ko didn't exist. I tried loading it
without the parameter, it loaded fine (but still doesn't work).
Also btw, I asked on the Demudi list whether anyone has this module
actually working. One user reported that he has the same problem, i.e.,
the module loads but is not working. Same kernel (2.6.14).
And thank you very much for helping me with this trouble.
Best,
dp
11:42 a.m.
Dave Phillips wrote:
The virtual MIDI ports convert between old programs using /dev/midi*
and new programs using ALSA sequencer ports. They do not connect
ALSA sequencer programs together; use the through ports for this.
I think you can achieve what you want by connecting a virmidi port and a
through port together.
HTH
Clemens
Clemens Ladisch wrote:
So the ALSA virtual MIDI ports are mapped to the raw MIDI ports ?
Yes. Anything sent to the sequencer ports will show up (only!) on the
raw MIDI ports, and vice versa.
Anything sent to a through port will be routed to all other connected
ports.
Please run the
following commands (in separate consoles):
aseqdump -p 80:0
amidi -p hw:2,0 -S "90 12 34 80 56 78"
or, for the other direction:
amidi -p hw:2,0 -d
aplaymidi -p 80:0 something.mid
Is there any output?
Yes, there is. But if I connect the virmidi port to a synth there's
still no sound.
12:06 p.m.
Clemens Ladisch wrote:
The virtual MIDI ports convert between old programs
using /dev/midi*
and new programs using ALSA sequencer ports. They do not connect
ALSA sequencer programs together; use the through ports for this.
I'm not sure I understand that. I've been connecting ALSA sequencer
clients with the virmidi ports for years. I can send a screenshot of my
MIDI connections in QJackCtl, with virmidi ports connected to QSynth,
Jack Rack, the Emu10k1 synth, etc. But I may misunderstand what you're
telling me.
I think you can achieve what you want by connecting a
virmidi port and a
through port together.
Been there, tried that, got no joy. :(
Btw, my results with this problem have been verified by another user of
Demudi and the 2.6.14 kernel. Should I be looking at anything else in
the system that might cause the trouble ? I'm terribly ignorant about
the 2.6 kernels...
Well, I can use the MIDI Thru for new work in Demudi.
Best,
dp
2:56 p.m.
Dave Phillips wrote:
Clemens Ladisch wrote:
Possibly grasping at straws.
I've just noticed under /proc/asound/VirMIDI is oss_mixer which when cat
gives zero levels.
Should there also be some sort of alsa mixer there? Or should the alsa
oss compatibility take care of that?
robin
The virtual MIDI ports convert between old
programs using /dev/midi*
and new programs using ALSA sequencer ports. They do not connect
ALSA sequencer programs together; use the through ports for this.
I'm not sure I understand that. I've been connecting ALSA sequencer
clients with the virmidi ports for years. I can send a screenshot of my
MIDI connections in QJackCtl, with virmidi ports connected to QSynth,
Jack Rack, the Emu10k1 synth, etc. But I may misunderstand what you're
telling me.
I think you can achieve what you want by
connecting a virmidi port and a
through port together.
Been there, tried that, got no joy. :(
Btw, my results with this problem have been verified by another user of
Demudi and the 2.6.14 kernel. Should I be looking at anything else in
the system that might cause the trouble ? I'm terribly ignorant about
the 2.6 kernels...
Well, I can use the MIDI Thru for new work in Demudi.
Best,
dp
19 Jan
19 Jan
4:49 a.m.
Dave Phillips wrote:
What are you actually trying to do?
Clemens Ladisch wrote:
A virmidi port has two ends: a rawmidi port and a sequencer port.
Anything written to the rawmidi device will be sent to all ports connnected
to the sequencer port, but cannot be read back from /dev/midi?.
Any messages arriving at the sequencer port can be read from the rawmidi
device, but are not sent to any other sequencer port.
The virtual MIDI ports convert between old
programs using /dev/midi*
and new programs using ALSA sequencer ports. They do not connect
ALSA sequencer programs together; use the through ports for this.
I'm not sure I understand that. I've been connecting ALSA sequencer
clients with the virmidi ports for years. I can send a screenshot of my
MIDI connections in QJackCtl, with virmidi ports connected to QSynth,
Jack Rack, the Emu10k1 synth, etc. But I may misunderstand what you're
telling me. I think you can
achieve what you want by connecting a virmidi port and a
through port together.
Been there, tried that, got no joy. :( From where to where to you want MIDI messages to be
routed?
Regards,
Clemens
7:42 a.m.
Clemens Ladisch escribe:
A virmidi port has two ends: a rawmidi port and a
sequencer port.
Anything written to the rawmidi device will be sent to all ports connnected
to the sequencer port, but cannot be read back from /dev/midi?.
Any messages arriving at the sequencer port can be read from the rawmidi
device, but are not sent to any other sequencer port.
Does this mean that a virmidi port just serves for sending copies of
MIDI data to different sequencers/sound generators?
Can I use vkeybd or a MIDI keyboard connected to the MIDI in of a MIDI
interface to play a virtual synthesizer without using virmidi?
Thanks for any comments.
Cordially, Ismael
--
Dropping science like when Galileo dropped his orange
8:40 a.m.
Ismael Valladolid Torres wrote:
Clemens Ladisch escribe:
The purpose of the snd-virmidi module is to serve as a bridge between old
programs that understand only /dev/midi* and new program that use the ALSA
sequencer API.
A virmidi port has two ends: a rawmidi port and a
sequencer port.
Anything written to the rawmidi device will be sent to all ports connnected
to the sequencer port, but cannot be read back from /dev/midi?.
Any messages arriving at the sequencer port can be read from the rawmidi
device, but are not sent to any other sequencer port.
Does this mean that a virmidi port just serves for sending copies of
MIDI data to different sequencers/sound generators? Can I use vkeybd or a MIDI keyboard connected to the
MIDI in of a MIDI
interface to play a virtual synthesizer without using virmidi?
Yes; vkeybd, the MIDI in, and the virtual synthesizer all use ALSA sequencer
ports.
HTH
Clemens
7:56 a.m.
Clemens Ladisch wrote:
What are you actually trying to do?
Dosemu has two methods for MIDI I/O. One is to use the midid daemon, but
I've had latency issues with it so I prefer the second method, which is
to create a link between ~/.dosemu/run/dosemu-midi and /dev/snd/CxDx.
Yes, I could route it to /dev/midi directly on my desktop machine
because it has a hardware MIDI port, but my laptop has no MIDI hardware.
Thus, if I want to run dosemu MIDI apps on the laptop I need the virmidi
module. The situation is nearly identical with Xsteem, I need the
virmidi module for running Xsteem on my laptop.
As I said before, if these applications abided by ALSA standards I might
be able to designate MIDI ports with the hw:x,x syntax. Alas, they
don't, and I can't.
Best,
dp
From where to where to you want MIDI messages to
be routed?
8:50 a.m.
Dave Phillips wrote:
Clemens Ladisch wrote:
Case 1: sending from Dosemu to an ALSA program:
- tell Dosemu to write to /dev/midi02;
- run "aseqdump" (or any other program);
- run "aconnect 80:0 aseqdump";
- data sent by a DOS program will be received by aseqdump.
Case 2: sending from an ALSA program to Dosemu:
- tell Dosemu to read from /dev/midi02;
- run "aplaymidi -p 80:0 something.mid",
or "aconnect 72:0 80:0" to route your external keyboard to Dosemu.
Do these work?
Regards,
Clemens
What are you actually trying to do?
From where to where to you want MIDI messages to be routed?
Dosemu [...] 
21 Jan
21 Jan
11:08 a.m.
Hallo,
Clemens Ladisch hat gesagt: // Clemens Ladisch wrote:
Case 1: sending from Dosemu to an ALSA program:
- tell Dosemu to write to /dev/midi02;
- run "aseqdump" (or any other program);
- run "aconnect 80:0 aseqdump";
- data sent by a DOS program will be received by aseqdump.
Case 2: sending from an ALSA program to Dosemu:
- tell Dosemu to read from /dev/midi02;
- run "aplaymidi -p 80:0 something.mid",
or "aconnect 72:0 80:0" to route your external keyboard to Dosemu.
What would be the best setup for a third case: sending from Dosemu to
Dosemu or more generally: sending from one program, that can only use
rawmidi/OSS ports to another program, that can only use rawmidi/OSS
ports?
Assuming I have a virmidi card as card 2, then I would have four ports
shown in aconnect: Virtual Raw MIDI 2-0', Virtual Raw MIDI 2-1' etc. I
know and actually used the fact, that these are available as rawmidi
ports in /dev/snd/midiC2D0, /dev/snd/midiC2D1, ...
So connecting like this "aconnect 80:0 80:1" will let me read at
/dev/snd/midiC2D1 what I write to /dev/snd/midiC2D0. Is there a way to
access these "sub-ports" through /dev/midi?? or /dev/amidi?? as well?
Ciao
--
Frank Barknecht _ ______footils.org_ __goto10.org__
6:25 p.m.
On Sat, 2006-01-21 at 16:08 +0100, Frank Barknecht wrote:
What would be the best setup for a third case: sending
from Dosemu to
Dosemu or more generally: sending from one program, that can only use
rawmidi/OSS ports to another program, that can only use rawmidi/OSS
ports?
Fix the apps?
;-)
Lee
6:46 p.m.
Hallo,
Lee Revell hat gesagt: // Lee Revell wrote:
On Sat, 2006-01-21 at 16:08 +0100, Frank Barknecht
wrote:
They aren't broken. ;)
In the past I just did some symlinks to /dev/snd/midiCxDx, which
works fine, but maybe this isn't necessary?
Ciao
--
Frank Barknecht _ ______footils.org_ __goto10.org__
What would be the best setup for a third case:
sending from Dosemu to
Dosemu or more generally: sending from one program, that can only use
rawmidi/OSS ports to another program, that can only use rawmidi/OSS
ports?
Fix the apps?
6:51 p.m.
On Sat, 2006-01-21 at 23:47 +0100, Frank Barknecht wrote:
Hallo,
Lee Revell hat gesagt: // Lee Revell wrote:
While many people think it's a neat trick, "cat foo > /dev/midi"
and /dev/dsp and ioctl()s on /dev nodes and what not is an interface we
are trying to move away from. So they are not necessarily broken, they
are just being stubbornly old fashioned.
Lee
On Sat, 2006-01-21 at 16:08 +0100, Frank
Barknecht wrote:
They aren't broken. ;)
In the past I just did some symlinks to /dev/snd/midiCxDx, which
works fine, but maybe this isn't necessary? What would be the best setup for a third case:
sending from Dosemu to
Dosemu or more generally: sending from one program, that can only use
rawmidi/OSS ports to another program, that can only use rawmidi/OSS
ports?
Fix the apps?
22 Jan
22 Jan
10:25 a.m.
Greetings:
First, thanks to everyone who has joined in on this topic. I'm not at
my desktop machine now, but I'll try Robin's symlink dodge when I get
home. Meanwhile, here's a link to a page that illustrates exactly what
it is that I want to achieve on Demudi. Hopefully it clarifies precisely
what I want to do:
http://linux-sound.org/virmidi-demo.html
Note that the system shown is running under Planet CCRMA, RH9, 2.4.26
low-latency kernel. It works beautifully on my 366 MHz PII laptop.
Best,
dp
10:59 a.m.
Hallo,
Dave Phillips hat gesagt: // Dave Phillips wrote:
First, thanks to everyone who has joined in on this
topic. I'm not at
my desktop machine now, but I'll try Robin's symlink dodge when I get
home. Meanwhile, here's a link to a page that illustrates exactly what
it is that I want to achieve on Demudi. Hopefully it clarifies precisely
what I want to do:
http://linux-sound.org/virmidi-demo.html
Okay, for this symlinks won't help, if your dosemu actually is writing
to /dev/snd/midiC1D0, as symlinks will just create an alias to that
device file.
You could try to "play" a midi file to /dev/snd/midiC1D0 with cat:
$ cat file.mid > /dev/snd/midiC1D0
while Qsynth is connected as shown in the screenshot. Again this
*should* work, but of course may not. *If* it works, then snd-virmidi
and the ALSA sequencer work as they should and I believe, there is
nothing broken in the Demudi kernel then.
Ciao
--
Frank Barknecht _ ______footils.org_ __goto10.org__
6921
days inactive
6926
days old
linux-audio-user@lists.linuxaudio.org
20 comments
6 participants
participants (6)
-
Clemens Ladisch -
Dave Phillips -
Frank Barknecht -
Ismael Valladolid Torres -
Lee Revell -
robin