Announcing ReaLearn: Sophisticated MIDI/MCU/OSC learn for REAPER

[foxAsteria]

Originally Posted by helgoboss

With this release, ReaLearn for the first time conquers REAPER's timeline (*Playtime glancing at ReaLearn slightly nervously*).

Can you give a few more details? I'm intrigued but not quite sure what you mean by this. Is this smooth seeking feature configurable by increments other than measures by chance?


I do have one small issue (more of a nitpick): when I filter by source or target and then use learn to change source or target, ReaLearn stays in filter mode but doesn't update for new input anymore until I hit stop and then filter again. Would be nice if it went right back to filtering. Not a huge deal tho.

Thanks helgo!

[helgoboss]

Originally Posted by foxAsteria

Can you give a few more details? I'm intrigued but not quite sure what you mean by this. Is this smooth seeking feature configurable by increments other than measures by chance?

The main novelty here - when compared to REAPER's built-in smooth-seek related actions - is feedback. This lets you use e.g. grid controllers like the Launchpad (or anything which has buttons with LEDs) to trigger regions with visual clue what's currently playing.

No, the smooth seek increments are not configurable. If this is something you would find useful, I could ask Justin if he can make something like this possible. But what do you have in mind exactly? UPDATE: BTW, the smooth seek increment is not just "end of measure" at the moment. When playing a region, it keeps playing until the end of the region.

[foxAsteria]

Originally Posted by helgoboss

If this is something you would find useful, I could ask Justin if he can make something like this possible. But what do you have in mind exactly?

Well typically in other DAWs, you can quantize the playhead based on any musical increment or the grid. Grid would be ideal.

Basically, it's performance related, The idea is to change between parts on the fly or to repeat them quickly for stutter effects etc. The smooth seek settings in Reaper are limited to whole measures, which makes it rather...uh, limited.

I suppose buffering a whole project to change position on the fly would be problematic at small increments, but I'd use it most often at 1/4 and 1/2 notes; never less than 16ths.

There are SWS actions for seeking regions which obey the seeking preferences (but again, they're only multiples of 1 measure). I do have a Launchpad X, and would totally love to have a custom page with realtime pads lit up for any regions in the project. But I take it this is only working with feedback for the native Reaper actions?

Man I think if you could somehow persuade Justin to allow measure divisions in the seek prefs, that would be amazing. It's must be 5 or more years by now since I made that request. I could never understand why it was like that in the first place.

[Wilson]

Hi, there is a Midi Feedback problem with the new ReaLearn update.
The Midi controller LED buttons are flickering like there is a Midi Feedback loop after starting Reaper.
I have LoopBe1 installed for a StreamDesk device and if I mute the LoopBe1 after starting Reaper and unmute again the Midi Feedback loop problem is gone.
If you can fix that that would be great! Thanks for your help!

[helgoboss]

Originally Posted by foxAsteria

Well typically in other DAWs, you can quantize the playhead based on any musical increment or the grid. Grid would be ideal.

Basically, it's performance related, The idea is to change between parts on the fly or to repeat them quickly for stutter effects etc. The smooth seek settings in Reaper are limited to whole measures, which makes it rather...uh, limited.

Ah yes, so you want it as in Playtime but for complete regions.

Originally Posted by foxAsteria

There are SWS actions for seeking regions which obey the seeking preferences (but again, they're only multiples of 1 measure). I do have a Launchpad X, and would totally love to have a custom page with realtime pads lit up for any regions in the project. But I take it this is only working with feedback for the native Reaper actions?

ReaLearn will switch on and off your LEDs no matter which technique (ReaLearn itself, SWS actions, native REAPER actions) you use to trigger the regions. It also will give you a light show if you just press play and the play cursor "drives" through the regions. That's because ReaLearn implements true feedback that is completely independent from control.

We could discuss ages here, so I would suggest to just download the new version and try it.

Originally Posted by foxAsteria

Man I think if you could somehow persuade Justin to allow measure divisions in the seek prefs, that would be amazing. It's must be 5 or more years by now since I made that request. I could never understand why it was like that in the first place.

Buffering time could be one reason why smaller time units are not available. Or - what I rather hope - it was just not implemented because there are 1000 other features in the queue to be implemented. I mean, this is probably for any software project in existance the number one reason why many things that are possible and make sense are not yet implemented

[helgoboss]

Originally Posted by Wilson

Hi, there is a Midi Feedback problem with the new ReaLearn update.
The Midi controller LED buttons are flickering like there is a Midi Feedback loop after starting Reaper.
I have LoopBe1 installed for a StreamDesk device and if I mute the LoopBe1 after starting Reaper and unmute again the Midi Feedback loop problem is gone.
If you can fix that that would be great! Thanks for your help!

Okay, go back to 2.5.0 in the meantime. I'm going to need the RPP. Please send to info@helgoboss.org.

[Wilson]

Originally Posted by helgoboss

Okay, go back to 2.5.0 in the meantime. I'm going to need the RPP. Please send to info@helgoboss.org.

ok I have send you the RPP.

Thanks a lot!

[nofish]

Originally Posted by foxAsteria

I do have one small issue (more of a nitpick): when I filter by source or target and then use learn to change source or target, ReaLearn stays in filter mode but doesn't update for new input anymore until I hit stop and then filter again. Would be nice if it went right back to filtering. Not a huge deal tho.

I noticed this too.

[helgoboss]

There can be errors with 2.6.0 when playing from the very start of the project or when using metronome count-in. Just go back to 2.5.0 for now if you are hitting these.

[Droinkus]

Originally Posted by helgoboss

Use these settings, they work for me: https://i.imgur.com/JRaifa3.png

Thanks for the quick response!


I still have the same issue where I can either receive messages;
Within Reaper's Preferences ("Listen") with no entry for OSC Pilot in ReaLearn
Or
In ReaLearn with no entry for OSC Pilot in Reaper's preferences
But not both simultaneously

If I receive messages in ReaLearn (with no entry for OSC pilot in Reaper's Preferences) I can receive messages via Learn, but the mappings do not work.

Here's my settings - perhaps I've just got something simple setup incorrectly?

https://imgur.com/a/XM82zFd

[helgoboss]

Originally Posted by Droinkus

I still have the same issue where I can either receive messages;
Within Reaper's Preferences ("Listen") with no entry for OSC Pilot in ReaLearn
Or
In ReaLearn with no entry for OSC Pilot in Reaper's preferences
But not both simultaneously

It's technically impossible that more than one "thing" listens to the same port. This is a characteristic of the UDP protocol on which OSC is based, not a ReaLearn or REAPER limitation.

If you really need both REAPER and ReaLearn connected to the same OSC/PILOT instance (why?), here are some potential approaches:

a) Maybe OSC/PILOT allows to send its messages to multiple hosts/ports?

b) Maybe there's some kind of OSC multiplexer that acts as a middleman and multiplexes the received OSC messages to multiple hosts/ports ... maybe OSCIIbot can do that?

Originally Posted by Droinkus

If I receive messages in ReaLearn (with no entry for OSC pilot in Reaper's Preferences) I can receive messages via Learn, but the mappings do not work.

Here's my settings - perhaps I've just got something simple setup incorrectly?

https://imgur.com/a/XM82zFd

The settings look okay (except that you can only have one of those, as mentioned above).

That your mappings don't work might be related to the specific OSC/PILOT template you are using. I don't own OSC/PILOT and the trial version can't load templates I think so first thing I would suggest you to try it with the default template (which just works for me). If this works for you, too, then you know it's related to the template.

Another OSC-related reason might be that you have selected the wrong OSC argument number in ReaLearn's mapping source.

Or it's something else which is not at all OSC related. In that case send your project file to info@helgoboss.org (please nothing complicated, then I can't look into it, just a minim project to reproduce the behavior).

[Droinkus]

Originally Posted by helgoboss

It's technically impossible that more than one "thing" listens to the same port. This is a characteristic of the UDP protocol on which OSC is based, not a ReaLearn or REAPER limitation.

If you really need both REAPER and ReaLearn connected to the same OSC/PILOT instance (why?), here are some potential approaches:

a) Maybe OSC/PILOT allows to send its messages to multiple hosts/ports?

b) Maybe there's some kind of OSC multiplexer that acts as a middleman and multiplexes the received OSC messages to multiple hosts/ports ... maybe OSCIIbot can do that?



The settings look okay (except that you can only have one of those, as mentioned above).

That your mappings don't work might be related to the specific OSC/PILOT template you are using. I don't own OSC/PILOT and the trial version can't load templates I think so first thing I would suggest you to try it with the default template (which just works for me). If this works for you, too, then you know it's related to the template.

Another OSC-related reason might be that you have selected the wrong OSC argument number in ReaLearn's mapping source.

Or it's something else which is not at all OSC related. In that case send your project file to info@helgoboss.org (please nothing complicated, then I can't look into it, just a minim project to reproduce the behavior).

For some reason as I was setting this up I thought I was only getting functionality with Touch OSC with both Reaper prefs and ReaLearn assigned.
Looks as though I'm incorrect.

I'll investigate as you've suggested...

Thanks - much appreciated!

[helgoboss]

Released ReaLearn 2.7.1 (ReaPack installation instructions)

Changes:
- #160 Added new track type "<Dynamic>" (allows to calculate a track number depending on ReaLearn's internal parameter values)
- #160 Improved target track picker GUI by making it easier to create ReaLearn presets that are going to be used across different projects (e.g. "By name" now lets you enter a string as one would expect)
- #160 Improved performance by trying to not send too much unnecessary feedback when mapping or mapping target activation states change
- #210 Improved handling of REAPER's "Audio device closed" state (in ReaLearn &lt; 2.5.0 being in that state for a long time could have led to slow memory exhaustion, in ReaLearn &lt; 2.7.0 it would eventually lead to an error message "Full" instead - which is already better, and in 2.7.0 it's handled as it probably should)
- #207 Fixed error message complaining about negative positions when using count-in and sometimes when playing from the very project start (introduced in 2.6.0)
- Improved usability by not intentionally generating a backtrace when pressing "Log debug info" (that caused confusion)
- Improved usability by reducing likelihood for audio dropouts by avoiding allocation in real-time thread when using very many mappings (probably premature optimization)
- Fixed another possible error message when closing a project while ReaLearn is open

Dynamic track calculation

I think the first change is noteworthy especially for the control surface fetishists among ReaLearn users. Some weeks ago, nofish basically asked if it's possible to have just one mapping and use it to control different tracks depending on which bank/program is currently selected. I was still unsure at that time if I want to add this feature ... well, now it is here. See the updated user guide for details (search for "<Dynamic>").

Before today, ReaLearn's internal parameters were only useful for the conditional activation concept - which is and will always be about switching mappings on and off, nothing more. That's great and you can achieve almost anything with this concept. However, there are scenarios in which you end up defining many mappings that only differ in the target track. In these scenarios, using conditional activation - although possible in theory - is just not practical. Starting from today, you need one mapping only whose target track is based on a position that you calculate from parameter values using a little mathematical expression, such as "p1 * 100".

I plan to bring the same possibility to other target elements during the next days:

- Track FX instance
- Track FX parameter
- Track send (send, receive and hardware output)

Current state

In a post some weeks ago, the question came up how ReaLearn compares to CSI. While I'm not familiar with CSI, I tried to list some of the things that I knew for sure ReaLearn couldn't do back then but CSI can:

Originally Posted by helgoboss

What can CSI do that ReaLearn can't?

1. OSC protocol (This is one of the next things that I want to add to ReaLearn. The concept is there already.)
2. Dynamic track positions, based on the current page number (see this post ... also something I plan to add and the concept is standing.)
3. AFAIK, this one can make it easier in some cases to use some hardware-specific features (Same here, I plan to add sys-ex support and custom source-specific settings.).
4. EUCON protocol (Mmh, I don't think I will add support for this, except someone brings up a really good argument. At first sight, this appears too exotic to me and too tailored to the typical control surface use cases).

Luckily, with the last bunch of ReaLearn updates this is not true anymore:

1. OSC: Now supported
2. Dynamic track positions: Now supported
3. MIDI sys-ex: Now supported (both for real-time feedback and at mapping activation/deactivation time)
4. EUCON: No plans to add that

[tompad]

Originally Posted by helgoboss

Released ReaLearn 2.7.1 (ReaPack installation instructions)

Changes:
- #160 Added new track type "<Dynamic>" (allows to calculate a track number depending on ReaLearn's internal parameter values)
- #160 Improved target track picker GUI by making it easier to create ReaLearn presets that are going to be used across different projects (e.g. "By name" now lets you enter a string as one would expect)
- #160 Improved performance by trying to not send too much unnecessary feedback when mapping or mapping target activation states change
- #210 Improved handling of REAPER's "Audio device closed" state (in ReaLearn &lt; 2.5.0 being in that state for a long time could have led to slow memory exhaustion, in ReaLearn &lt; 2.7.0 it would eventually lead to an error message "Full" instead - which is already better, and in 2.7.0 it's handled as it probably should)
- #207 Fixed error message complaining about negative positions when using count-in and sometimes when playing from the very project start (introduced in 2.6.0)
- Improved usability by not intentionally generating a backtrace when pressing "Log debug info" (that caused confusion)
- Improved usability by reducing likelihood for audio dropouts by avoiding allocation in real-time thread when using very many mappings (probably premature optimization)
- Fixed another possible error message when closing a project while ReaLearn is open

Dynamic track calculation

I think the first change is noteworthy especially for the control surface fetishists among ReaLearn users. Some weeks ago, nofish basically asked if it's possible to have just one mapping and use it to control different tracks depending on which bank/program is currently selected. I was still unsure at that time if I want to add this feature ... well, now it is here. See the updated user guide for details (search for "<Dynamic>").

Before today, ReaLearn's internal parameters were only useful for the conditional activation concept - which is and will always be about switching mappings on and off, nothing more. That's great and you can achieve almost anything with this concept. However, there are scenarios in which you end up defining many mappings that only differ in the target track. In these scenarios, using conditional activation - although possible in theory - is just not practical. Starting from today, you need one mapping only whose target track is based on a position that you calculate from parameter values using a little mathematical expression, such as "p1 * 100".

I plan to bring the same possibility to other target elements during the next days:

- Track FX instance
- Track FX parameter
- Track send (send, receive and hardware output)

Current state

In a post some weeks ago, the question came up how ReaLearn compares to CSI. While I'm not familiar with CSI, I tried to list some of the things that I knew for sure ReaLearn couldn't do back then but CSI can:



Luckily, with the last bunch of ReaLearn updates this is not true anymore:

1. OSC: Now supported
2. Dynamic track positions: Now supported
3. MIDI sys-ex: Now supported (both for real-time feedback and at mapping activation/deactivation time)
4. EUCON: No plans to add that

I see a light in the tunnel - and its NOT the 12 o'clock train comming!
:-)

[Stroudy]

Originally Posted by helgoboss

I could implement that, it's possible to access that name. There could be a second entry in the "Auto-load FX preset" dropdown named "Depending on focused FX and preset". I knew this dropdown would be good for something

Oh cool. Thanks Helgo!!

[X-Raym]

@helgoboss
Hi !
We have selected track as Target, but is there Selected Tracks (plural) ? :P
For track volume for eg.


Thx!

[helgoboss]

Originally Posted by X-Raym

@helgoboss
Hi !
We have selected track as Target, but is there Selected Tracks (plural) ? :P
For track volume for eg.


Thx!

Do you mean this? https://github.com/helgoboss/realearn/issues/4. If yes, then the answer is no, but you can upvote this FR by pressing the smiley.

However, if all you want to do is to select multiple tracks at once by the press of one button, you can create one mapping with "Track selection" target for each track to be selected, both mapped to the same button.

Or if you want just one mapping and you have a folder track and want all tracks in this folder to get selected at once, then let me know. I was thinking to support this use case.

[vonglan]

Hi,
does anyone here have any experience with the Akai APC40Mk2 and ReaLearn? I don't yet have an LED-ring controller that makes use of ReaLearn's feedback, and wonder how the Akai compares to the Firefighter Twister. Are the 16 LED-Ring knobs functionally equivalent to what the Twister offers? (I found the "communications protocol" document, but it was not really clear for me.)
Thanks,
Edo

[foxAsteria]

Man, I thought ReaLearn was the ticket to finally get some use out of TouchOSC encoders, but I simply can't get them to work at all. No matter the settings, clockwise maxes the parameter and counter-clockwise minimizes it completely; nothing in-between.

Anyone successfully using TouchOSC encoders in Reaper? What settings?

[lowellben]

Originally Posted by foxAsteria

Man, I thought ReaLearn was the ticket to finally get some use out of TouchOSC encoders, but I simply can't get them to work at all. No matter the settings, clockwise maxes the parameter and counter-clockwise minimizes it completely; nothing in-between.

Anyone successfully using TouchOSC encoders in Reaper? What settings?

Same. I can't use TouchOSC at all with RL.

[foxAsteria]

Originally Posted by lowellben

Same. I can't use TouchOSC at all with RL.

Oh well I mean everything else is working great, just not encoders. TouchOSC has kind of a steep learning curve but I finally got it working after some years of trying now and again. Even better now with ReaLearn.

I plan to post some tips and templates in the relatively near future. ReaLearn with make it easy to share, with all the action mappings already done.

Are you running TouchOSC bridge? If you've set up MIDI in your templates, you need that to send it to Reaper and ReaLearn. It'll give you some new MIDI ports for that purpose.

[helgoboss]

Originally Posted by foxAsteria

Man, I thought ReaLearn was the ticket to finally get some use out of TouchOSC encoders, but I simply can't get them to work at all. No matter the settings, clockwise maxes the parameter and counter-clockwise minimizes it completely; nothing in-between.

Anyone successfully using TouchOSC encoders in Reaper? What settings?

I do: https://i.imgur.com/H6fT5I1.png (using the built-in "Jog-On" template, left jog wheel in this example). What template do you use?

Originally Posted by lowellben

Same. I can't use TouchOSC at all with RL.

Need more info. Error message, anything?

Does it say "unable to connect"? Then make sure REAPER itself doesn't have an OSC control surface defined that uses the same ports.

Originally Posted by foxAsteria

Oh well I mean everything else is working great, just not encoders. TouchOSC has kind of a steep learning curve but I finally got it working after some years of trying now and again. Even better now with ReaLearn.

I plan to post some tips and templates in the relatively near future. ReaLearn with make it easy to share, with all the action mappings already done.

Are you running TouchOSC bridge? If you've set up MIDI in your templates, you need that to send it to Reaper and ReaLearn. It'll give you some new MIDI ports for that purpose.

Haven't used the TouchOSC bridge so far. I thought it's only necessary if the application doesn't support OSC and you have to stick to MIDI instead?

[foxAsteria]

Originally Posted by helgoboss

I do: https://i.imgur.com/H6fT5I1.png (using the built-in "Jog-On" template, left jog wheel in this example). What template do you use?

I made my own templates. Thanks, I'll try your settings here and see if anything's different.

I believe OSC is also over-network-only, so Bridge would be needed to send MIDI over usb.

[helgoboss]

Originally Posted by foxAsteria

I made my own templates. Thanks, I'll try your settings here and see if anything's different.

I believe OSC is also over-network-only, so Bridge would be needed to send MIDI over usb.

The important thing is the "Is relative" checkbox.

[Droinkus]

Is Type Int something that might be supported in the future?

It looks as though OSC Pilot only sends Ints for Buttons and the Z Axis on the XY pad.

OSC Pilot widgets that have Floats work fine.
The only way I can get functionality from a Button is to not have any argument - which works well for some things I need to do, but I can't all the functionality from Toggles.

I'm using MIDI at the moment for toggles but I'd like to migrate to full OSC with ReaLearn.

Thanks!

[helgoboss]

Originally Posted by Droinkus

Is Type Int something that might be supported in the future?

It looks as though OSC Pilot only sends Ints for Buttons and the Z Axis on the XY pad.

OSC Pilot widgets that have Floats work fine.
The only way I can get functionality from a Button is to not have any argument - which works well for some things I need to do, but I can't all the functionality from Toggles.

I'm using MIDI at the moment for toggles but I'd like to migrate to full OSC with ReaLearn.

Thanks!

Can you find out what Int values OSC Pilot sends when using the XY pad and which when using the buttons? What's the value range? The only reason why I just ignore Int at the moment is that ReaLearn needs to know about the possible value range (min - max) to make sense of the incoming values. With float and double this is usually by convention a normalized value (min = 0.0, max = 1.0), so it's straightforward to support it. With bool it's dead easy because there are only 2 possible values (off/on). But with Int I don't think there's any convention what's the minimum and the maximum value. I guess I should provide a little text field which lets you enter the minimum/maximum values.

So yes, I could add that feature! With your help.

[Droinkus]

Originally Posted by helgoboss

Can you find out what Int values OSC Pilot sends when using the XY pad and which when using the buttons? What's the value range? The only reason why I just ignore Int at the moment is that ReaLearn needs to know about the possible value range (min - max) to make sense of the incoming values. With float and double this is usually by convention a normalized value (min = 0.0, max = 1.0), so it's straightforward to support it. With bool it's dead easy because there are only 2 possible values (off/on). But with Int I don't think there's any convention what's the minimum and the maximum value. I guess I should provide a little text field which lets you enter the minimum/maximum values.

So yes, I could add that feature! With your help.

Sure of course happy to help - ReaLearn is an integral part of my workflow!

OSC Pilot sends ints of either 1 or 0 .
This is true of both the Z axis on the XY Pad and Buttons.

The bug (hopefully not feature) I've found in OSC Pilot is that initially Floats are output, and ReaLearn works great, but after closing and opening, Buttons and the Z axis output ints of 1 or 0.
I've posted on the OSC Pilot forum, we'll see what I learn from there.

[helgoboss]

Released ReaLearn 2.8.0-pre1 (ReaPack installation instructions)

Changes:
- #212 Added new FX instance selector "Dynamic"
- #213 Added new FX parameter selectors "Dynamic" and "By name"
- #214 Added new send/receive/output selectors "Dynamic", "By ID" (source/destination track ID) and "By name"
- #72 Added option to use track receives (change of perspective) and hardware output sends instead of track sends only
- #217 Added possibility to use wildcards (\* and ?) in all name selectors to match only parts of names
- #218 Fixed error when controlling "Last touched" target after the last touched target is gone (e.g. after track deleted)
- #208 Fixed too early rotation when using "Incremental buttons" in combination with "Rotate"

With these changes, most target "objects" (tracks, effects, effect parameters, sends/receives) support all of the following ways of referring to them:

- Dynamic (dynamic selection based on parameter values)
- By ID (sticky reference that survives reordering/insertion/removal)
- By position
- By name (exact or wildcard pattern match)

Would be glad if some of you can test this and give feedback. I tested it with my projects but a greater test coverage would be nice for this kind of changes, that's why I published it as a prerelease.
(Prereleases must be selected manually in ReaPack using right mouse click on the ReaLearn pack => Versions).

[foxAsteria]

Originally Posted by helgoboss

(Prereleases must be selected manually in ReaPack using right mouse click on the ReaLearn pack => Versions).

Can't seem to find that anywhere.

EDIT nvm you have to go to "browse packages" and search there. I've never seen that window before. I guess because I don't use the default menus.

[helgoboss]

Originally Posted by Droinkus

Sure of course happy to help - ReaLearn is an integral part of my workflow!

OSC Pilot sends ints of either 1 or 0 .
This is true of both the Z axis on the XY Pad and Buttons.

The bug (hopefully not feature) I've found in OSC Pilot is that initially Floats are output, and ReaLearn works great, but after closing and opening, Buttons and the Z axis output ints of 1 or 0.
I've posted on the OSC Pilot forum, we'll see what I learn from there.

That sounds like a bug in OSC Pilot indeed. For buttons, 0 and 1 makes sense somehow but for a Z axis not really. Let's see what they respond.

[pyrannial]

Wow! This software is incredible, thank you so much! The whole world should be raving about it (well, at least the music production world).

[Droinkus]

Originally Posted by helgoboss

That sounds like a bug in OSC Pilot indeed. For buttons, 0 and 1 makes sense somehow but for a Z axis not really. Let's see what they respond.

Luckily there's an easy fix (another "Duh" on my part):

Just enter 0.0 and 1.0 in OSC Pilot and the output remains as Float after restart.

Here's the thread:
https://forum.oscpilot.com/t/floats-...ter-save/371/4


Again, thanks for your support and for getting OSC in ReaLearn!

[helgoboss]

Originally Posted by pyrannial

Wow! This software is incredible, thank you so much! The whole world should be raving about it (well, at least the music production world).

Thanks, glad you see it like that Spreading the word would be helpful to make more people know about this.

Originally Posted by Droinkus

Luckily there's an easy fix (another "Duh" on my part):

Just enter 0.0 and 1.0 in OSC Pilot and the output remains as Float after restart.

Here's the thread:
https://forum.oscpilot.com/t/floats-...ter-save/371/4


Again, thanks for your support and for getting OSC in ReaLearn!

He raises a point there with Ints being something that should be handled. But he sees it from a perspective which doesn't care about "meaning" of the particular OSC messages, just about "transmitting". It would be not good to assume that Int is the same like Bool because it has the capacity to express much more than just on and off. I think at some point (if somebody really needs it or this comes up more often) I will provide an additional field where one can enter the expected value range of the data type (in your case it would be 0 to 1).

[helgoboss]

Released ReaLearn 2.8.0-pre2 (ReaPack installation instructions)

Changes:
- #166 Added "Fire after timeout" mode for button presses
- #32 Added "Fire after timeout, keep firing" mode for button presses (turbo mode!)
- #223 Added "Seek" target (including feedback on beat changes or in high resolution!)
- #225 Fixed bug in "Go to marger/region" target that caused markers/regions to be triggered even on button release

After the somewhat tedious work of the last days (bringing all target objects like tracks, sends, FX parameters etc. on the same level), today I was doing myself the favor of implementing some "fun" mode and target additions. It was a rewarding experience because these additions are quite interesting, but they were relatively easy to implement, no bad surprises like "Oh no, now I have to rebuild 20% of the plug-in in order to support this". Looks like ReaLearn's internal design gets a lot of things right. Really glad I did the rewrite last year.

But enough developer musings ... enjoy the additions and let the light show begin (or motorized fader show or whatever)

[earlabs]

hi

i just updated from 1.10 to 2.71.
1.10 I ran together with LBX_stripper and everything was hunky dory (with MFR). After upgrade some knobs don't respond anymore so I guess I'll have to do my settings again. Takes some work but that's okay.

What puzzles me, though, is why on earth I cannot add any mapping when I choose auto-load preset, depending on focused FX.

In my setup (with 1.10) I could switch between various limiters, compressors, eq's and whatnot and MFR knob settings followed. But now I'm empty handed.

What am I missing?

[helgoboss]

Originally Posted by earlabs

hi

i just updated from 1.10 to 2.71.
1.10 I ran together with LBX_stripper and everything was hunky dory (with MFR). After upgrade some knobs don't respond anymore so I guess I'll have to do my settings again. Takes some work but that's okay.

Oh, I tried hard to make things backward compatible but maybe something slipped through.

Originally Posted by earlabs

What puzzles me, though, is why on earth I cannot add any mapping when I choose auto-load preset, depending on focused FX.

In my setup (with 1.10) I could switch between various limiters, compressors, eq's and whatnot and MFR knob settings followed. But now I'm empty handed.

What am I missing?

You are not missing anything, it's designed that way. If you switch an instance to "Auto-load preset", this instance will do nothing else than automatically loading presets. You are not at all empty handed though: Just fire up a second instance and use it for whatever you want to do.

There's nothing that you could do in 1.10 that you cannot do in 2.x.

[foxAsteria]

Originally Posted by earlabs

\
What am I missing?

Took me a minute to get my head around too; had like 4 instances all doing the same thing before I realized only one plugin can be focused at once and so all the presets can be handled by the same auto-loader instance (but that's all it will do after that).

If you want to edit presets from that instance though, you have to temporarily turn off auto-loading and select the preset you want to edit manually.

[earlabs]

I finally used the function of adding many instances. That helped a lot. And, and I don't know if that is the way it is supposed to do, I did a group for each vst (equalizer, limiter, imager, etc).

Somehow all my older projects are fucked up, though. It's not too bad, because most are 'done' and I only need to work with my mouse instead of the MFR (which is an awesome machine, btw).

[helgoboss]

Originally Posted by earlabs

I finally used the function of adding many instances. That helped a lot.

Genuine question: Any idea of what I could do to effectively make users more aware of the fact that ReaLearn can (and should) be instantiated multiple times? It's difficult for me to put myself into the shoes of new users after years of developing this thing haha. But it always striked me as obvious that a VST can be opened multiple times. Like e.g. if I have a synth, I can add multiple instances of it, why not.

Originally Posted by earlabs

And, and I don't know if that is the way it is supposed to do, I did a group for each vst (equalizer, limiter, imager, etc).

Yes, the question of "best practice" comes up a lot. It really depends on what you want. On ReaLearn's issue tracker there are currently some discussions which also touch this question and I hope that soon I will be able to provide recommendations which practice will be best for which kind of use case. This documentation FR I just created contains some basic hints about multiple instance usage already: https://github.com/helgoboss/realearn/issues/236

Somehow all my older projects are fucked up, though. It's not too bad, because most are 'done' and I only need to work with my mouse instead of the MFR (which is an awesome machine, btw).

This is unexpected. Can you provide details? Best thing would be to send me the RPP file which still used the old version (info@helgoboss.org) along with a small hint which settings exactly were fucked up when migrating to 2.71.

[earlabs]

I am a visual thinker, so I find the large amounts of text too much. Perhaps a step-by-step explainer would help. With sort of schematics. I can help you work something out.