RapidComposer v4 beta feedback and discussion (locked)

[lulukom]

Hi Attila. I think, after all, quantization of midi controllers does not work.

[sj1]

Hi.

While the topic is current, I'd like to ask what exactly is meant by "quantization of midi controllers"?

I ask from both the user perspective (describe what you'd like to see happen, in terms that make both musical and programming sense), and from the developer perspective (i.e. what is the implementation/algorithm approach in RC 4.7b1).

IOW, beginning with a bunch of notes and CC data that have been free-recorded off-the-grid, what is supposed to (or does) happen to the CC data (between where and where) when the notes are quantized to the grid?

IOW, describe the algorithm.

[BluGenes]

Well, I had to have a little fun with your question, so, I entered a modified version of the question into ChatGPT and this is the results..






CC messages are separate from Notes, two different things. So, lining them up correctly is the issue..

[lulukom]

I am very surprised not understanding the problem ) . If you've played a solo with a pitch controller and then quantized the notes but didn't quantize the pitch controller, you probably won't recognize your solo ? Try it and maybe you will understand what we are fighting for here?

[sj1]

Magic = the desired expression goes with the (moved) notes. Sure. But that's not an algorithm. It was the algorithm I'm asking about.

There are typically many CC messages during and in-between the notes, and as recorded, they are "continuous". Once you decide that as-recorded continuity is no longer what you want to keep, there is a real question about what to create in its place, is there not?

[sj1]

Will messages be eliminated? Will "fill-in" messages be created? Will extraneous messages that had no aural effect on the original playback be identified and discarded? Will there be millisecond windows around the start and end of notes during which messages are considered to "go with" that note? I'm probably not even scratching the surface. I'm asking as-if I had to implement it in code (retired programmer here). IMO, the goal(s) are not only difficult to implement, but difficult to define. But if both are done successfully and w/o error then the Magic will be ours to use! <g>

[sj1]

"Any sufficiently advanced technology is indistinguishable from magic." ― Arthur C. Clarke

[lulukom]

Yes , you are right that of course I am not a programmer and I do not know what algorithms make musical maneuvers in the program . But I only know that I have a DAW in which I have been working for 30 years)) and she at least tries to do it in such a way that it suits me (well, in most cases)) and I haven’t thought about it for a long time ... But I had to think about the RC ... And of course, you described the possible approaches to algorithms in a very interesting way.

[sj1]

Cubase?

[lulukom]

MOTU Digital Performer 11.21

[BluGenes]

Yo, Atilla, Can you set the chord key change graphic thingy to be user defined color setting? Or, oh man, this is reaching though.. lol, color coded.. I could use green for this part and maybe purple for that part, etc.. just an idea anyways..

[sj1]

Here is a future wish:

Add a wizard that would analyze whatever chords are in the Master Track and generate a new Chord Rules Set (CRS). The user could then further edit the CRS to expand it out.

One contemplated use would be to type or drag in chords from any desired source (e.g. a known tune), generate a CRS, try out using that CRS and then add to it to expand it (either manually, or with another merged import) until it is working well. At that point you have a new, cool custom CRS derived from a specific genealogy. Rinse and repeat, growing your CRS library.

This would work especially quickly in combination with MusicXML chord import.

Here is a further implementation suggestion re: the above -

Given a desirable future situation where a CRS analysis of MT chords can be used to either start a new CRS or add to an existing CRS, at each point when new rules are added to a CRS assign that set of rules a sequence number (seqno).

IOW, all the rules added to a CRS upon it's first creation (or save) are given a seqno of 1. The next time we add rules to that CRS, those new rules get a seqno of 2, etc.

Then, as a CRS grows to include ever larger seqnos, at any later point we could "reach back" and throw out all the rules with a specified prior seqno.

Example: I am building a CRS out of "romantic ballads". I've added 10 of them, generating 10 seqnos. Then I decide it's all getting too tame, so I add chords from a heavy metal tune and a country tune (hey, it's an experiment!). Rules from these are tagged as CRS seqnos 11 and 12.

Well, let's say I judge that experiment a failure. I want to be able to throw out rules from seqno 11, test with rules 1-10, and 12. Then judge that, and either keep 12 or throw it out too and move on.

IOW, every set of rules added to a CRS is considered an "atom" and the CRS itself is a molecule. Atoms would retain their identity and be able to be removed (or even added) independently thereafter.

IOW, no information loss, unless explicitly flushed by the user.

Programmatically this would be relatively straightforward I think. Musically? Well, that will be revealed when users start to work with it.

FWIW, I do think users would be able to run with this capability to create some very unique and useful CRSs.

[sj1]

Hi,

I'd like to offer a MIDI import example and comment on it.

I've attached the .mid file as a .zip.

Here is the chord sheet for it:

26-2.png

and here is a pic of the chords RC chose for the first 4 bars:

RC - 26-2 01.jpg

In general, I would say the matchup of analyzed chords is not very good.
Specifically for example, I would wonder how RC could possibly classify the first chord as having a minor 3 or minor 7.

So, comment #1 is: Master Track chord analysis for importing from MIDI files may need some attention.

Comment #2 is: allowing Master Track chord importing from MusicXML files would completely solve this sort of thing (for cases where a MusicXML file containing the chords is available).

[ebbownee]

I recently purchased the full version of RapidComposer. How can I participant int he beta?

[BluGenes]

I recently purchased the full version of RapidComposer. How can I participant int he beta?

Use the contact support info in RC.. Just mention an interest and you should be able to get set up..

Once you are set up, it is best to keep the beta install separate from the production install. Just put the beta in a different install directory.. (just add a beta directory to the install directory is easiest)

Also, there is no way of separating the production VST from the beta VST, so, which ever was installed last will be used in your DAW.