Textual tempos play back wrong in time signatures other than X/4
To reproduce:
1. In a recent nightly build (post-#82466: Add back textual tempo texts), add the Andante tempo text to a score in cut time (¢, or 2/2).
2. Press play.
Expected result: tempo is half note = 92.
Actual result: tempo is quarter note = 92, or half note = 46.
The tempo is off by other multiples in compound time signatures.
This is more or less a regression, because this was an issue in MuseScore 1 (Create > Text > Tempo...), is no longer in MuseScore 2 (Add > Text > Tempo being "smart" about the time signature), and has made a comeback in the MuseScore 3 nightlies.
Comments
It is not really a regression, 2.0 simply didn't have textual tempo texts, and the tempo texts it has are not smart about the time sig at all, they just show exactly what you get (quarter note = 80, half note = 80, etc).
But consider Add > Text > Tempo—that's smart enough to have the various alternatives lined up and fit with the time signature accurately, isn't it?
If youhave a 4/4 time sig, and add a 'quarter note = 80' and then decide to change timesig to 2/2, where is that tempo text smart? If it were smart, it'd change to `half note =80', and that is is what you expect from an 'Andante'
That's another matter, which indeed definitely isn't a regression.
Well, no, if you add quarter note = 80, that's exactly what you get - quarter note at 80. Therer are no "smarts" here, nothing that used to work but now longer does. Hence, no regression.
What we would need is to add some *new* smarts so that Andante doesn't use quarter note = 80 necessarily but rather "beat" = 80, whwere the interpretation of beat depends on tempo. And like I have said before, this is just not possible in the general case - even in 4/4, Andante might mean *very* different things depending on whether the piece basically has a motor rhythm involving quarter notes, eighth notes, or sixteenths. We can certain make better guess than assuming quartyer note = 80, but there is no getting around the fact that these markings are not absolutely - they are subjective depending on context as well as the performer. So, we make a reasonable stab at a guess, let the user edit.
See also https://musescore.org/en/node/82466#comment-640071, for the place(s) where a possibly usefull algorithm to calculate the tempos might be taken from.
If so, hopefully in a way that could and would make it into 2.1 too
According to the discussion, it is more like a feature request, than a bug. I propose to change title to something like "Support smart tempo in cut time signatures"