Closed rigaux closed 1 month ago
I will do this work soon.
Hi @rigaux, I'm not calling this done by any means, but on the develop branch, l have moved lyrics from the GeneralNotesOnly
detail level to the AllObjects
detail level. So if you choose GeneralNotesOnly
, you will not get lyrics, and if you choose AllObjects
(or AllObjectsWithStyle
, or AllObjectsWith...
) you will get lyrics along with all the key signatures, dynamics, metronome marks, etc.
Hopefully that is helpful in the short term; I'll work on the finer-grained command line options using this issue.
That's fine. Thanks a lot.
OK, on the develop branch now, you will find a brand-new command line that allows you to specify exactly which details you would like to have compared. For example:
python3 -m musicdiff -d GeneralNotes Lyrics Style -- file1.musicxml file2.krn
(I have not made this as fine-grained as you might have wanted; please list the specific things you would like that are missing: e.g. beams, directions, etc)
On the develop branch (and soon to be released as musicdiff v4.0.0) the command line is much more fine-grained:
python3 -m musicdiff -i decoratednotesandrests lyrics style -x beams -- file1.musicxml file2.krn
arguments:
-i/--include one or more named details to include in comparison (the default is allobjects,
a.k.a. decoratednotesandrests and otherobjects). Can be decoratednotesandrests,
otherobjects, allobjects, or any combination of those and/or the following:
notesandrests; the aforementioned note decorations: beams, tremolos, ornaments,
articulations, ties, slurs; the other objects: signatures, directions,
barlines, staffdetails, chordsymbols, ottavas, arpeggios, and lyrics; and
a final few details that are not found in allobjects: style, metadata, and
voicing. voicing compares how notes are included in voices and chords (by
default this is ignored).
-x/--exclude one or more named details to exclude from comparison. Can be any of the
named details accepted by -i/--include.
-o/--output one or both of two output formats: text (or t) or visual (or v); the default
is visual). visual (or v) requests production of marked-up score PDFs; text
(or t) requests production of diff-like text output.
file1 first music score file to compare (any format music21 or converter21 can parse)
file2 second music score file to compare (any format music21 or converter21 can parse)
I just released musicdiff v4.
Great !
Hi We are working on an OMR system which focuses (until now) on pure music aspects. Even though we plan to include lyrics in a near future, it would be nice to add a detail level that excludes lyrics (sth like "GeneralNotesNoLyrics".
Also, since the number of "with/without" combinations is likely to seriously augment the number of levels, maybe consider instead one option per feature on the command line: --no-lyrics=1 --no-beams=1 --no-directions=1, etc.
Thanks for the great work!
Philippe