Automatic MIDI generation on repeat

[tkna91]

It would be useful to have an option to generate MIDI on repeat.
You can check the MIDI content listening and stop playback if you like it, and if you don't stop, the MIDI will be regenerated on repeat.
You can find great melodies and rhythms just by leaving it alone.
I think a very big advantage is that you can leave it alone and concentrate on the listening.

Screenshot:

[RobU23]

Hi !

Thank you for your request, it's a great idea. I am not sure that this will be implemented in the current version - the codebase is quite fragile and unmanagable (this is what happens when you create a program while learning the language (Lua), and an API (Reaper). The script is non-realtime and has no concept of time in relation to the song position.

I have been working on a new version that has a cleaner more modular codebase, and perhaps your idea may be possible to implement ... but, my free time to work on this is very limited at present so for the moment, the script is what it is..

Again, thanks for your input, it's nice to know that people are still using the script :)

Regards
RobU

[tkna91]

If the Randomise/Generate button is externalized as an action, it may be possible to achieve this with a marker action. This would be less time-consuming to implement and more scalable.
https://www.sws-extension.org/markeractions.php

[sconte19]

I’ve stumbled upon the exact same problem: I’m doing a CS bachelor thesis in procedurally generated adaptive audio music for videogames.

In practice I've used MIDI_EX_MACHINA on Reaper + Realearn to use OSC to communicate with unity - it works surprisingly well.

The problem is that I wanted to generate MIDI at runtime (Which i had to drop) and that would’ve opened a plethora of possibilities.

The proposed original system was a refactoring of MIDI_EX_MACHINA to instead run programmatically by “filling” a certain midi item (specifically using track number + midi_item number to make a 2D map) - I actually managed to somewhat make it work, but it doesn't function properly 1:1 and thus I had to scratch it and just use it manually for the demo.

After having that function it would’ve just been a matter of having a configuration file somewhere with all the generation instruction for each take (ideally generated via a GUI like a html file that fills a json) + a reader function in reaper that would accordingly generate and move Reaper to the correct region.

@RobU23 , given your expertise, this refactor from GUI to programmatic should be straightforward—would you collaborate to complete it? Full credits to you, of course - Think about it the project would allow with just an OSC controller to generate music in real time +
modulate any parameter using REALEARN - the possibilities would be insane.

I’d be more than willing to showcase the project, let me know if you’d be interested - I’d like to share this tool with everyone, giving Reaper users another great tool on their belt - I’m particularly interested in the applications for live music.