On Wed, Jul 3, 2024 at 11:07 PM Werner LEMBERG <[email protected]> wrote:
> > OK, here are some scans that I've found. As you can see, this kind of > subdivision is not so special as previously assumed. > > * Jehan Alain, trois danses – Deuils (for organ), (publisher: Leduc) > > * Prokofiev, piano sonata 7, op. 83, second movement (publisher: > Muzgiz) > > * Rachmaninov, piano concerto 3, op. 30, third movement (publisher: > Gutheil) > > It looks to me like there are two different issues here: 1) subdividing the lowest level of notes in groups of 4, not groups of 2. All three examples show this. IIUC, this could be solved in our current code by eliminating some levels of subdivision (we currently do it for every power of 2, instead of for "every other" power of 2). 2) Using a nonstandard number of beams between the subdivided parts. The Prokofiev piece uses a standard number of beams (when grouping 4 of the shortest notes, there should be two fewer -- not one fewer -- beams between the groups, according to the theory). The Alain and Rachmaninov only have one fewer beam, so the beam count between the groups is not appropriate for the lengths of the subdivided groups, according to the Gould rules. Personally, I think the Gould rules are correct, but two of the music publishers in your example do not agree with me. And they should probably have MUCH more credibility than I do. At any rate, I think it's important to understand both issues. When I made my music function, I was only responding to item 2, not to item 1. Thanks, Carl
