Survey: Would you use a plugin to automate orchestral expression?

[sica]

Hi everyone,

I'm a software developer and hobbyist composer. I spend way too much time drawing CC curves for my orchestral tracks, and I'm exploring whether there's a better solution.

The idea is a VST plugin that analyzes your MIDI performance (melody contour, dynamics, harmonic tension, phrase structure) and generates intelligent starting points for expression curves and articulation changes. Think of it as an assistant that handles the tedious groundwork, leaving you free to focus on the creative refinements. Essentially, cutting CC programming time from hours to minutes using (possibly) AI/ML.

I would ask you few quick questions:

1. How much time do you normally spend on CC automation per track?
2. If so, what bother you the most?
3. What features would be essential vs. nice-to-have
4. Would you pay for a tool that automates this? If so, what price seems fair?

I'm aware that other tools exist, but there seems to be room for something more.

Early exploration phase, no code yet. Your feedback determines if this is worth building.

Thanks for any insights!

EDIT: I've added a poll. Please feel free to vote and also share more detailed thoughts in the comments - qualitative feedback is just as valuable as the numbers

[Aloysius]

Yes.

1. How much time do you normally spend on CC automation per track? Not much. Too time consuming / fiddily.
3. What features would be essential vs. nice-to-have? Load up a instrument in Kontakt and have the expression work without having any need to map things in Kontakt.
4. Would you pay for a tool that automates this? If so, what price seems fair? The cheaper the better. It's not something I'd use very often but it would be very nice to have.

[sica]

Yes.

1. How much time do you normally spend on CC automation per track? Not much. Too time consuming / fiddily.
3. What features would be essential vs. nice-to-have? Load up a instrument in Kontakt and have the expression work without having any need to map things in Kontakt.
4. Would you pay for a tool that automates this? If so, what price seems fair? The cheaper the better. It's not something I'd use very often but it would be very nice to have.

Thank you for your feedback!

[Aloysius]

I did my best. I hope others will chime in for you.

[jamcat]

I use SWAM instruments. I'd be thrilled with a plugin that could generate realistic performance articulations for it, particularly for the strings.

[sica]

I use SWAM instruments. I'd be thrilled with a plugin that could generate realistic performance articulations for it, particularly for the strings.

Cool! Can you explain how you would like the tool to be used? What features you would expect from it?

[jamcat]

I think ease of use would be my top priority. Maybe something that provides a set of articulation keyswitches which then outputs the CC necessary to create the selected articulation. It could perhaps have some limited controls for things like speed, intensity, and random variation, to fine tune it within a narrow range, but the results should always be useable and user control shouldn’t bog it down or complicate its use. The idea is to uncomplicate SWAM and provide good, realistic articulations with ease every time.

I would use such a plugin to quickly create a workable foundation of realistic performance automation data in my DAW, which I could refine by hand later if needed.

The goal of such a plugin should be to bridge the gap between easy-to-use keyswitches as found in sample libraries and the infinite performance variations available through parameter automation in SWAM.

I would like to be able to just stamp out a bunch of spiccato strokes, for example, without having to figure out what that would look like then draw all the automation data in by hand, or perform it in real-time with a midi keyboard and expression controllers. SWAM has a lot of parameters for shaping the performance other than just expression, and a combination of moves are typically required for a realistic performance.

[sica]

I think ease of use would be my top priority. Maybe something that provides a set of articulation keyswitches which then outputs the CC necessary to create the selected articulation. It could perhaps have some limited controls for things like speed, intensity, and random variation, to fine tune it within a narrow range, but the results should always be useable and user control shouldn’t bog it down or complicate its use. The idea is to uncomplicate SWAM and provide good, realistic articulations with ease every time.

I would use such a plugin to quickly create a workable foundation of realistic performance automation data in my DAW, which I could refine by hand later if needed.

The goal of such a plugin should be to bridge the gap between easy-to-use keyswitches as found in sample libraries and the infinite performance variations available through parameter automation in SWAM.

I would like to be able to just stamp out a bunch of spiccato strokes, for example, without having to figure out what that would look like then draw all the automation data in by hand, or perform it in real-time with a midi keyboard and expression controllers. SWAM has a lot of parameters for shaping the performance other than just expression, and a combination of moves are typically required for a realistic performance.

Thank you! I will consider your suggestions to better shape the idea.

[Ah_Dziz]

I wouldn't be particularly interested. The majority of sample libraries don't respond well enough for a generalized set of contours and best guess from input cc data. I can guess on my own and by the time I get the hang of a library I can outperform something that's guessing. For libraries that do have the depth to properly respond (swam is the only one really) I've been using a breath controller for ages now and it very much speeds things up to the point where something like the proposed plugin would be too much. I did have this issue with layering a single performance across multiple sample libraries / instruments in unison to generate enough randomness (time, vibrato depth, playing intensity) for it to sound like separate instruments. I was eventually able to get it done on my own and my little tool does fine.

[planetearth]

This sounds interesting, but how would you make sure it could do what you want it to do with all (or at least, most) of the different "orchestra" plug-ins available? Some respond to expression using MIDI CC 11, but some (such as EastWest Quantum Leap) use CC 1 or CC 7 for "expression" control. (Yes, yes, we all know it's wrong. But it doesn't stop them. ) Even Kontakt isn't consistent, depending upon the library and what the original developers designed it to do.

Some orchestra (or orchestra section, or solo instrument) plug-ins also have different expression responses. For example, huge, multi-sampled, "pro" libraries have (in some cases) every note or every other note sampled at multiple velocities. Obviously, smaller libraries have fewer samples. In each case, the difference between the relatively limited range of MIDI CC 11 values may be more obvious.

I don't know if you were considering it, but hardware samplers/ROMplers would also have different—and, probably, compressed—responses to CC 11 for the same reason smaller software plug-in libraries do: memory conservation. But some people (myself included) turn to older samplers for orchestra sounds when working on projects designed to sound like something from the late '80s or early '90s to '00s.

And that's just CC 11. Obviously, other MIDI CC controls would face similar issues. Kontakt, EastWest, IK Multimedia, Spitfire, Vienna Symphonic Library, etc. all do things slightly differently, so it might be difficult to get the same results out of them if you (as the developer) don't have all these libraries to test.

And how would the plug-in deal with the MIDI 2.0 spec and all the changes it brings to MIDI CC control, values, and per-note options?

I like the idea of something like this. I'm just not sure how it would work across plug-ins and libraries. But you've probably already considered that, and I look forward to seeing how this progresses!

Steve