patches dont sound like they should, as soon as floorboard finishes loading.
both 2021 and 2022 version have this issue.
tried on different laptops / my PC (diff versions of linux , but
reasonably up to date code )
all same result.
==> on windows 10, same laptop, it works with no issues....
==> on a windows inside KVM virtualization on linux, .... it works with
no issues.. strange, because the virtual machine uses USB from the linux
host....
It seems to be an issue like not being able to communicate properly over
usb :
tones hanging, tones spontaneous sounding.... (funny, since we
cannot use gr55 as a controlled synthesizer with e.g. a midi keyboard ? )
even simple things like note-off do not work.
I think patches dont get initialized or only partly initialized with
all parameters; parameters are missing (thats what it sounds like )
I m also using ardour, a audio / midi studio recording / mixer software.
ardour does record the outgoing midi from gr55, the notes are all there,
GR55 itself sounds good, ardour is not influencing it, no problems.
( i use the midi with musescore to see what i have been playing while
improvising :-) )
so apparently, it is the configuring via midi in of the patches that
does no longer work properly.
on the GR55, I changed the USB MIDI setting "switch" to ON.
this is necessary if you want to record the midi OUT stream
that apparently upsets the GR55 Floorboard software: on Linux
things work normally on Windows 10 ( same hardware )
it is NOT the USB stack in linux : If I run a windows 10 virtual on top of same linux,
things work as they do on Windows running natively on same hardware.
My conclusion is, that something is different in
GR55 Floorboard for linux vs Floorboard for Windows.
I hope it can be fixed.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Both Windows and Linux (and Mac OS too) all use exactly the same source code, so there are no internal differences in GR-55FloorBoard between platforms except for platform compilation.
One thing that will kill midi communication is if the operating system midi routing has a loopback somewhere, sent data the GR-55 has output is put into its midi input (You can tell if the GR-55 becomes a bit sluggish on patch change).
Check any midi routing in Jack/ALSA or in your DAW application, that it is not sending unnecessary midi data back to the GR-55.
For the false note triggering, that can only be as a result of the USB cable ground causing a noisy earth loop with the GR-55 and creating digital noise into the GK guitar input.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
while testing I disabled (shut down) ardour. thats the only midi aware software I have.
I do not use Jack or similar, too much complexity.
Alsa / Pulse, thats all.
Anyway, it sounds like that could be it anyway,
I will check and let know
thanks for the suggestion.
Last edit: rensg 2023-06-12
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
same source code is not a guarantee for identical behaviour.
as soon as I/O is involved, it is definitely not guaranteed to work the same way.
multiplatform dev tools like Lazarus might go a long way, but not the whole way.
Floorboard example:
e.g: on windows, I can still select patch by using the GK controller on the guitar (up/down buttons). Floorboard Windows does not notice.
On linux, this will not work. patch selection can only be done through Floorboard linux,
while floorboard is active. The up/down bottom on the GK does nothing.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
FWIW on bare metal my desktop PCs run Linux only. While I have got several Windows version running as guests in virtual machines and while I also have got the Linux branch of gr-55_floorboard installed, I only used windows branches running on wine.
I'm not up to date, at the moment the installed Linux version is 2020-4. By different wine prefixes two versions are installed, 2018-4 and 2020-4.
At the moment I'm on
$ pacman -Q wine-staging
wine-staging 8.9.1-1
Over the time I used several versions of wine and wine-staging. IIRC only one update years ago has broken the GUI, but that could be solved by downgrading wine.
To be honest, I usually use the GR-55 just for MIDI, only in the beginning I used its synth and modeling, hence it's a while back that I used gr-55_floorboard a lot. However, that time wine worked very good and it much likely still works today.
When on Linux consider to run the Windows branch on wine.
For me testing the Linux branch was just a waste of time.
Regards,
Ralf
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
FWIW when I tested the Linux branch a long time ago the app behaved different to the Windows branch.
I can't comment on the OP's MIDI issues. Indeed USB MIDI is a PITA related to galvanic isolation and jitter. To get rid of extra jitter caused by USB I'm in favour of PCI/PCIe cards. However, to get galvanic isolation, even an USB audio device can be used, just don't connect two devices directly via USB, use 5 pin DIN connectors. An audio device, with audio IOs and 5 pin DIN MIDI connectors should use optocouplers for the MIDI signal and should not connect ground from one device with another, IOW it should grant galvanic isolation.
Regards,
Ralf
Last edit: Ralf Mardorf 2023-06-11
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
tried a bunch of scenario's
on same hardware ( 8 core, 16GB, ssd disks, absolute realtime prio, unlim pinned memory )
with midi data sending out turned on:
via USB : windows works, linux does not; have to turn off midi sending, only then it works fine.
via DIN, into a USB audio interface box: both work normally on windows and linux
windows floorboard on wine : works, but initial load of all user patches is very slow compared to windows or linux on hardware.
in a midi monitor, I can see that GR55 device generates tons of pitch bend events going OUT when on USB, but much less on DIN.... it just keeps doing that, I m not even touching the instrument...
Accuracy of the midi stream as recorded is much better on DIN than on USB.
nothing goes INTO the GR55, apart from when I write MIDI with floorboard.
seems to me the GR55 synth is killing itself.. when using USB.
conclusion is I will have to hook it up with DIN into a USB midi/interface.
or plug in the USB for working the patches, and DIN cable midi-out when I m recording.
thanks a lot to you all for all suggestions / ideas / input.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
noticing that GR55 device apparently has an issue with MIDI over USB...
I bought an extra 5pin-MIDI I/O <=> USB interface
Bus 005 Device 004: ID 7104:1400 CME (Central Music Co.) U2MIDI Pro
for 26 Euro
which seems to work just fine.
I can now record midi out and control patches with Floorboard without issues and no cable plugging / unplugging needed.
response is quick, smooth. MIDI recording is accurate.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Great to see a positive result, using guitar to midi and the editing software at the same time does not produce desirable results at times, due to both using the same midi I/O stream. The data between real-time playing and large flows of patch system exclusive data become mixed together and also cause midi traffic congestion.
I would recommend keeping the midi out transcribing, and patch editing as separate activities (switch off the Guitar to midi while editing).
Cheers :-)
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
yes, absolutely: i turn midi out OFF, when editing patches.
by the way : there is a "THIN DATA" setting, which discards most of the pitch bend events.
unless one needs them, I suggest turning that on... it saves a lot of messaging that will not end up in the notation anyway.
note: using a hub is not a good idea.... I use USB socket directly connected to motherboard..
Last edit: rensg 2023-06-26
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
patches dont sound like they should, as soon as floorboard finishes loading.
both 2021 and 2022 version have this issue.
tried on different laptops / my PC (diff versions of linux , but
reasonably up to date code )
all same result.
==> on windows 10, same laptop, it works with no issues....
==> on a windows inside KVM virtualization on linux, .... it works with
no issues.. strange, because the virtual machine uses USB from the linux
host....
It seems to be an issue like not being able to communicate properly over
usb :
tones hanging, tones spontaneous sounding.... (funny, since we
cannot use gr55 as a controlled synthesizer with e.g. a midi keyboard ? )
even simple things like note-off do not work.
I think patches dont get initialized or only partly initialized with
all parameters; parameters are missing (thats what it sounds like )
I m also using ardour, a audio / midi studio recording / mixer software.
ardour does record the outgoing midi from gr55, the notes are all there,
GR55 itself sounds good, ardour is not influencing it, no problems.
( i use the midi with musescore to see what i have been playing while
improvising :-) )
so apparently, it is the configuring via midi in of the patches that
does no longer work properly.
I found out what the "problem" is ...
on the GR55, I changed the USB MIDI setting "switch" to ON.
this is necessary if you want to record the midi OUT stream
that apparently upsets the GR55 Floorboard software: on Linux
things work normally on Windows 10 ( same hardware )
it is NOT the USB stack in linux : If I run a windows 10 virtual on top of same linux,
things work as they do on Windows running natively on same hardware.
My conclusion is, that something is different in
GR55 Floorboard for linux vs Floorboard for Windows.
I hope it can be fixed.
Both Windows and Linux (and Mac OS too) all use exactly the same source code, so there are no internal differences in GR-55FloorBoard between platforms except for platform compilation.
One thing that will kill midi communication is if the operating system midi routing has a loopback somewhere, sent data the GR-55 has output is put into its midi input (You can tell if the GR-55 becomes a bit sluggish on patch change).
Check any midi routing in Jack/ALSA or in your DAW application, that it is not sending unnecessary midi data back to the GR-55.
For the false note triggering, that can only be as a result of the USB cable ground causing a noisy earth loop with the GR-55 and creating digital noise into the GK guitar input.
while testing I disabled (shut down) ardour. thats the only midi aware software I have.
I do not use Jack or similar, too much complexity.
Alsa / Pulse, thats all.
Anyway, it sounds like that could be it anyway,
I will check and let know
thanks for the suggestion.
Last edit: rensg 2023-06-12
same source code is not a guarantee for identical behaviour.
as soon as I/O is involved, it is definitely not guaranteed to work the same way.
multiplatform dev tools like Lazarus might go a long way, but not the whole way.
Floorboard example:
e.g: on windows, I can still select patch by using the GK controller on the guitar (up/down buttons). Floorboard Windows does not notice.
On linux, this will not work. patch selection can only be done through Floorboard linux,
while floorboard is active. The up/down bottom on the GK does nothing.
Hi,
FWIW on bare metal my desktop PCs run Linux only. While I have got several Windows version running as guests in virtual machines and while I also have got the Linux branch of gr-55_floorboard installed, I only used windows branches running on wine.
I'm not up to date, at the moment the installed Linux version is 2020-4. By different wine prefixes two versions are installed, 2018-4 and 2020-4.
At the moment I'm on
$ pacman -Q wine-staging
wine-staging 8.9.1-1
Over the time I used several versions of wine and wine-staging. IIRC only one update years ago has broken the GUI, but that could be solved by downgrading wine.
To be honest, I usually use the GR-55 just for MIDI, only in the beginning I used its synth and modeling, hence it's a while back that I used gr-55_floorboard a lot. However, that time wine worked very good and it much likely still works today.
When on Linux consider to run the Windows branch on wine.
For me testing the Linux branch was just a waste of time.
Regards,
Ralf
Hi,
FWIW when I tested the Linux branch a long time ago the app behaved different to the Windows branch.
I can't comment on the OP's MIDI issues. Indeed USB MIDI is a PITA related to galvanic isolation and jitter. To get rid of extra jitter caused by USB I'm in favour of PCI/PCIe cards. However, to get galvanic isolation, even an USB audio device can be used, just don't connect two devices directly via USB, use 5 pin DIN connectors. An audio device, with audio IOs and 5 pin DIN MIDI connectors should use optocouplers for the MIDI signal and should not connect ground from one device with another, IOW it should grant galvanic isolation.
Regards,
Ralf
Last edit: Ralf Mardorf 2023-06-11
tried a bunch of scenario's
on same hardware ( 8 core, 16GB, ssd disks, absolute realtime prio, unlim pinned memory )
with midi data sending out turned on:
via USB : windows works, linux does not; have to turn off midi sending, only then it works fine.
via DIN, into a USB audio interface box: both work normally on windows and linux
windows floorboard on wine : works, but initial load of all user patches is very slow compared to windows or linux on hardware.
in a midi monitor, I can see that GR55 device generates tons of pitch bend events going OUT when on USB, but much less on DIN.... it just keeps doing that, I m not even touching the instrument...
Accuracy of the midi stream as recorded is much better on DIN than on USB.
nothing goes INTO the GR55, apart from when I write MIDI with floorboard.
seems to me the GR55 synth is killing itself.. when using USB.
conclusion is I will have to hook it up with DIN into a USB midi/interface.
or plug in the USB for working the patches, and DIN cable midi-out when I m recording.
thanks a lot to you all for all suggestions / ideas / input.
just to finish up this item :
noticing that GR55 device apparently has an issue with MIDI over USB...
I bought an extra 5pin-MIDI I/O <=> USB interface
Bus 005 Device 004: ID 7104:1400 CME (Central Music Co.) U2MIDI Pro
for 26 Euro
which seems to work just fine.
I can now record midi out and control patches with Floorboard without issues and no cable plugging / unplugging needed.
response is quick, smooth. MIDI recording is accurate.
Great to see a positive result, using guitar to midi and the editing software at the same time does not produce desirable results at times, due to both using the same midi I/O stream. The data between real-time playing and large flows of patch system exclusive data become mixed together and also cause midi traffic congestion.
I would recommend keeping the midi out transcribing, and patch editing as separate activities (switch off the Guitar to midi while editing).
Cheers :-)
yes, absolutely: i turn midi out OFF, when editing patches.
by the way : there is a "THIN DATA" setting, which discards most of the pitch bend events.
unless one needs them, I suggest turning that on... it saves a lot of messaging that will not end up in the notation anyway.
note: using a hub is not a good idea.... I use USB socket directly connected to motherboard..
Last edit: rensg 2023-06-26