[MusicXML] Title converted to copyright on import if positioned at bottom of page
When I open the attached MXL file with the 2.0 beta, the title text somehow becomes part of the canvas, on every page.
As such, not only can it not be edited (since it's not a text element anymore), it also can't be moved (since it's not even a vector graphic). It's simply stuck there in the background behind the score.
The file was an export from the same version of MuseScore.
Attachment | Size |
---|---|
March.mxl | 12.34 KB |
Comments
FWIW, it appears this text is represented in the MusicXMl file in a way that makes it appear to be copyright info. So you can edit it via File / Info, or suppress its display using Style / General / Header, Footer, Numbers. If you think it should not have been exported as copyright info, please post the original score from before you exported it to MusicXML so we can understand why it was turned into copyright.
Absolutely. It's actually the same score I attached to #42236: MSCZ score's text style settings in 1.3 distorted in 2.0 — Score.mscz . So, first the placement of the text elements is messed up when the MSCZ is opened in the beta, then when it's exported to MXL and opened in the beta again, the title has become the copyright?
The offsets in the styles are not expected by the MusicXML exporter in MuseScore 1.3, which results in incorrect y positions for the title page texts:
"Untitled" at y = 106 (bottom of the page)
"Composer and Composer" at y = 2959 (top of page)
"March" at y = 1481 (middle)
As the MusicXML importer puts text above the middle of the page in a vbox at the top, and text below the middle of the page in the copyright, this results in a vbox containing the title "Composer and Composer" and a copyright "Untitled March".
Given the way MusicXML and MuseScore are designed, this is the best algorithm I could come up with. Feel free to suggest improvements in how MuseScore 2.0 works (1.3 will not be changed anymore). Note that changes to the text placement algorithm on MusicXML import will have to work "well enough" for a wide range of cases, not just for your case.
Well, if we're updating the title, might as well make it more precise.
Just to be clear, exporting the same score to MusicXML from MuseScore 1.3 and opening it with MuseScore 1.3 had its own problems with messing up those text elements' layout, though nothing like this.
Thanks, Marc, File/Info is exactly what I needed. I have a bunch of PDFs with set numbering at the bottom center of the form:
- R7 -
and these get converted to copyright upon import as well.
( sorry if this is not the right place to post given the "[MusicXML]" tag ... )
If I'm not mistaken, it's actually the same thing—I believe the experimental PDF converter (at https://musescore.com/import) converts first to MusicXML, and then converts that to MSCZ.
Having checked MuseScores behaviour with a large set of MusicXML sample files, I concluded that converting credit-words (used by MusicXML for printed text such as title, composer and rights) to copyright metadata based on the current heuristics hurts more than it helps.
My current opinion is that it is best to set the copyright metadata only using the MusicXML rights element. I'll change the MusicXML importer accordingly.
Pull request 2528
Fixed in branch master, commit 32e36c4e9c
fix #42281 - [MusicXML] Title converted to copyright on import if positioned at bottom of page
Fixed in branch master, commit 64f633bf61
Merge pull request #2528 from lvinken/musicxmlfixes
fix #42281 - [MusicXML] Title converted to copyright on import if pos…
Fixed in branch 2.0.4, commit 2a071e4eb3
fix #42281 - [MusicXML] Title converted to copyright on import if positioned at bottom of page
Automatically closed -- issue fixed for 2 weeks with no activity.