Hive2 Filter Cutoff: spring-loaded synth joystick, sound returns to lowest point of cutoff, not the start point

[glokraw]

edit: this is solved 99%, see next post below


In anticipation of the new rythym soundset, I've been trying a lot of the existing Hive loops,
and when releasing the spring-centering joystick on my Kawai synth, the sounds don't re-center.
Many sounds have the up-ward bend mapped to the modwheel, and it shows in Hive as playing the horizontal x-axis 2, which is often controlling Filter Cutoff.

In the preset panel, the sound dials all seem to start at 12:oclock, moving the joystick upwards to engage the modwheel sends the Filter Cutoff to the lowest point of the cutoff with the slightest nudge, rather than starting the sound where the preset has it. Then, releasing the joystick doesn't return it to the starting 12:oclock setting, but to the lowest point of the cutoff, where many are pretty quiet, and some are nearly silent, until you reload or choose
another preset. This seems to hold true across sound types, and in the linux Hive, as well as the windows version in the AVLinux default wine-staging 6.22, and using the newest Reaper .669 editions..
It's an old Kawai, maybe that's how synth joysticks behaved in the ice age?

So I tried sounds in ZebraHZ that had cutoff on the modwheel, and they audibly seemed fine with returning to center. The old Firebird synth sounds also re-centered OK. (The horizontals axis are pitchbend on the Kawai, and seemed accurate in the testing. It's in pretty good shape overall.)

[glokraw]

MX Linux guru dolphin_oracle quickly posted a solution, such that the xy pad is now dealing accurately with Hive2 presets. His text file hides the gamepad from the linux system, but not
from wine and reaper, and there's no odd start/end positions when triggering the various modwheel sounds.

----------------------------------------------------------------------------------------------------------------------

make a file

Code: Select all

/etc/X11/xorg.conf.d/50-joystick.conf

with the following contents

Code: Select all

Section "InputClass"
Identifier "joystick catchall"
MatchIsJoystick "on"
MatchDevicePath "/dev/input/event*"
Driver "joystick"
Option "StartKeysEnabled" "False"
Option "StartMouseEnabled" "False"
EndSection

and see after a login/logout if the gamepad works without moving the mouse.

credits here:

https://spacebums.co.uk/mx-linux-gaming-tweaks/

[tasmaniandevil]

Hi, glad you found a solution.

But I'm not sure I fully understand the initial problem.

moving the joystick upwards to engage the modwheel sends the Filter Cutoff to the lowest point of the cutoff with the slightest nudge, rather than starting the sound where the preset has it. Then, releasing the joystick doesn't return it to the starting 12:oclock setting, but to the lowest point of the cutoff

The mod wheel is a unipolar parameter ranging from 0 to 127. So it's quite logical that it won't jump back to the middle position but to its lowest value if controlled via a bipolar joystick, if the joystick ranges from -64 to +63 or -128 to +127.

Many sounds have the up-ward bend mapped to the modwheel, and it shows in Hive as playing the horizontal x-axis 2, which is often controlling Filter Cutoff.

This is hard to understand.
If the mod wheel does control the XY, then you most likely have the mod wheel assigned to it in the MIDI learn menu.

Or could you name a preset from the factory library that exhibits the mentioned issue?

[glokraw]

I'll preface this by saying Howard's sounds are great!

And the screens below are from an unmodified AVLinux, where the gamepad is still 'a mouse', and the joystick behavior in the pics is from a Kawai synth that has a spring-centered joystick, the synth is connected with an E-mu X-midi usb interface.

The first pic is default setting of preset HS Jive-to-the-hive (in the untuned-loops section.) After a midi-learn on both XY 1 and 2, using the down-bend of the Kawai joystick and just nudging the joystick, sends both 1 and 2 to their extreme positions. (note the related dials)

I would expect that the sound would begin where the preset shows it, (leaving aside definitions) and that a nudge would slightly modify settings, and not invoke a jump to any extreme position first, and that the sound would stick to the joystick coordinates, and that the 1 and 2 icons would re-center when the joystick is released.

I think this would be more crucial on lead sounds in the hands of a skilled player.

How does this compare to Mac/Win behavior with a synth that has a spring-centered joystick?
D50, Wavestation, TG33, prophet, or Erica Black Joystick 2? I'm sure the testing offices must have them all somewhere

Jive-to-the-hive-default.jpg

nudged-joystick.png

The Hive and Zebra sounds are such treasure chests, I'm thinking that a dedicated tool for mapping gamepads and arrow keys would be a saleable product. Midi learn is pretty cumbersome in most software synths, and then under-utilized, yet the capability is pretty awesome, and could be highlighted by Kinesis, Kling Klong, and no-doubt the new Zebra Legacy collection.

(pics are from the Hive 2 "Swarm" skin, if folks are interested, it's a great gui, as is the original.)
Cheers

[glokraw]

Interesting somewhat related page

https://synthandsoftware.com/2021/08/jo ... s-game-on/

[tasmaniandevil]

Ah, thanks, understanding this better now.

I would expect that the sound would begin where the preset shows it, (leaving aside definitions) and that a nudge would slightly modify settings, and not invoke a jump to any extreme position first

When you start moving the joystick it starts sending values and the XY reacts to that value.

When you assign a CC to a parameter, it ranges from 0 to 127, so the XY starts with 0 at the left edge, and reaches the right edge at 127.

If the stick starts with 0 in the middle position, then this would explain why the XY jumps to the leftmost position when you start moving the stick.

Do you know how the joystick works, i.e. which CC value range it sends going from center to right or from center to left? That would be interesting.

Or can you say which Kawai synth you are using, then we could have a look at the manual to find out how exactly the joystick is sending CC data.

[glokraw]

Here's a link to the owners manual.

https://kawaius.com/wp-content/uploads/ ... Manual.pdf

http://www.synthark.org/Kawai/PH50.html

Pretty sure it's a derivative of the Kawai K1, as the sounds are compatible, but with limited editing.

[tasmaniandevil]

Ah, thanks for the link. Unfortunately the manual contains no details about the data the joystick is sending.

[glokraw]

Switching midi input to more modern Alesis QS8, which has a cluster of four midi-learnable sliders, things are more predictable. Starting with the four sliders in their lowest position, I load a preset,
'AS Pollination' in this case, which plays normally.

In both the linux and win-in-wine Hives, the AS Pollination preset is silenced by nudging slider 2, which is assigned to the filters, and jumps to the far left of their range. Quite a few sounds seem to have this mapping, and receive the similar filter-position silence...

linux-Hive-alesis-slider-Pollination preset silenced by filters.png

win-Hive-in wine alesis slider AS Pollination preset slider-2-nudged.png

BUT... in the win-Hive version in wine-staging, using the Logitech gamepad, there is no jumping! The action is smooth as silk, in all directions, and on both of the joypad control sticks.

Alesis QS8 manual:

https://www.manualsbase.com/manual/2244 ... lesis/qs8/

Kawai K1 manual, (PH 50's big brother synth):

https://kawaius.com/wp-content/uploads/ ... Manual.pdf

(as an aside, 8 of the gamepad buttons are mapped to a chromatic scale, could be fun for someone...especially if a latch is available, not sure how much is there inside Reapers joystick scripts, which have few lines of code to my untrained eye, which are found somewhere around:

.wine/drive_c/users/Reaper/InstallData/Data/joystick_midi

or wherever Reaper is installed.

Cheers