Am 17.01.26 um 18:00 schrieb Richard Shann:
On Sat, 2026-01-17 at 16:13 +0000, berenbeor wrote:
[...]
I am trying to do something that I think is quite simple, but after
hours of trying many programs can't do it. Basically clean up OCR PDF
piano scores. I created MXL files with audiveris, but the output has
a lot of mistakes, and I just want to do very simple cleanup, e.g.
move notes, add dropped notes, etc.
As you probably know, most of the other programs (Musescore, Dorico,
LMMS etc) prioritize measure time rules over notes, which is really
hard to comprehend. If I try to add a missing note, it often deletes
a correct note. I can't understand why time rules take precedence
over notes. Someone could make a bunch of money by creating a simple
program that allowed a user to turn off 'time rules' for a measure or
the score, insert whatever note/duration desired, and then go back
and fix the timing (perhaps with a simple toggle that shows the
measures needing fixing turning red). The idea of forcing random
notes and rests into a measure for a time rule is simply not how
people think.
Yes, changing your input while you are putting it in as if at each step
you have a complete finished score is annoying - I remember thinking
that when I gave musescore a go some years back. Denemo sort of does
the opposite: although it displays barlines to break the input into
chunks for the convenience of the display they are not necessarily
where the barlines will be, LilyPond determines that (unless you
override it). Instead Denemo just colors under and overfull bars blue
and red.
Thinking along these lines, currently there are Denemo measures that
divide the score into chunks and finally Lilypond measures (barlines) in
the typeset score. The Denemo measures seem quite artificial to me, as
they do not influence the final output. I know they can be adjusted with
Measures > Split Measure at Cursor and Measures > Merge with Next
Measure, which I use quite often. Nevertheless, it would be much more
intuitive for the user if Denemo measures would always adjust themselves
automatically while editing, just as Lilypond measures do. Especially
when arranging or transcribing by ear, that would save quite some extra
work fiddling with the artificial Denemo measures. What do you think
about refactoring the Denemo measures to automatically adjust themselves?
Andreas
