Implement collision avoidance between rests and notes/rests in other voices
Reported version
3.0
Priority
P1 - High
Type
Functional
Frequency
Few
Severity
S5 - Suggestion
Reproducibility
Always
Status
closed
Regression
No
Workaround
Yes
Project
OS: Windows 10 (10.0), Arch.: x86_64, MuseScore version (64-bit): 3.0.0, revision: 3543170
Repro steps
- Create a new score
- Press F to enter a quarter F
- Select all and exchange voice 1 and 2 back and forth
- Select the measure rest in voice 2 and hit 0 to enter a quarter rest in voice 2
Expected: The quarter rest is moved down to avoid collision
Actual: The quarter rest collides with the note
Workaround: Move the rest manually
| Attachment | Size |
|---|---|
| Collision-note-rest-voices.mscz | 2.34 KB |
Fix version
3.5.0
Comments
Different voices don't participate in collision detection for now.
This also affects imported 2.x scores. IMV, it's a bug.
In reply to This also affects imported 2… by geetar
I vote bug.
Well, it was never designed to work, never worked in any previous release, so it's not a bug in the usual sense of the word. It's simply a feature that has never been implemented. Obviously, though, it's one people might reasonably expect to see, and it's a very good idea to implement some day, regardless of what one calls it. So, it's a high priority suggestion. Please don't change these fields.
If it had worked in a previous release it would be a regression. This is not a criteria for bug or not. For example the not selectable instrument names have been working in 2.x but are declared as designed in 3.0, not bug.
But the main feature of MS 3.0 is automatic placement of all sort of fancy staff. In this view a not working automatic placement of such a basic thing as a rest is a bug, IMO.
+1. This should be extended to imported 2.x scores as well.
https://github.com/musescore/MuseScore/pull/5897/commits
Do you know what Elaine Gould says about this (in her book)?
Fill me in
Avoiding collisions is always good, of course. The only question would be about exactly how to avoid them - how much to move things, which things need to be moved in addition to remain aligned, etc. For example, in the original example here, simply moving the quarter rest on beat 1 might not be ideal, then it wouldn't align with the other rests in that measure. But maybe sometimes that's OK. There would really be a need to consult different engraving experts, published music, etc, to understand what's ideal. Gould doesn't go into a lot of detail here, but on p. 36 does suggest that you align rests that represent part of the same beat, also try to align them with other notes in that voice. It definitely looks to be more an art than a science.
Chen Lung thankfully pointed me to some pages in Behind bars
The current implementation may be not perfect. But it solves 95% of my scores automatically. Especially when toggling concert pitch with large transpositions it proved very helpful. And of course the automatic avoidance always can get manual corrections.
I can add some improvements according to the above list.
Fixed in branch 3.x, commit 61c680813a
fix #278538: Implement collision avoidance between rests and notes (if measure rest, for the whole measure) in opposite voices
Fixed in branch 3.x, commit 13b6db227a
fix #278538: Fine tuning in collision avoidance between rests and notes to pass the vtests + collect_artifacts
Fixed in branch 3.x, commit 68df2b2ad4
_Merge pull request #5897 from tobik/278538-restcollisionavoidance
fix #278538: Implement collision avoidance between rests and notes_
Automatically closed -- issue fixed for 2 weeks with no activity.