strange behaviour of dynamics lines
In Ornstein, Leo: Piano Sonata No.4, SO 360: complete.mscz, measures 35-36. This is what I expect should happen:
1. The existing dynamic is "mf" (m31), so the crescendos from (m31-m34) are ignored
2. The decrescendo hairpin of m35 should change the dynamic from "mf" (preceding) to "p" (succeeding).
3. The "dim." of m35 should have no effect ("p" -> "p" == no change)
4. The decrescendo hairpin of m36 should have no effect, since the preceding dynamic is "p" (of the same measure), followed by "f" (m40)
5. Measure 36, and measure 37 up till the "cresc." should remain "p", after which there should be a gradual crescendo till measure 40.
This is what happens:
1. Measure 35: both hairpin decrescendo and "decr." are ignored
2. Measure 36: sudden decrease from "mf" to "p" at "p"
3. Measure 36: hairpin decrescendo ignored (expected)
4. Measure 36: sudden increase from "p" to "f" on the third beat, right after the hairpin
5. Mesaure 37-40: the gradual crescendo has no effect ("f" -> "f" == no change)
Why?
Comments
The crescendo and decrescendo markings expect both a dynamic preceding and a dynamic following. Right now, you have a crescendo, followed by two decrescendos, and only the last one has an actual dynamic attached to the end. You'll have to stick in some invisible dynamics to get the behavior that you want. (De)crescendos without dynamics behave oddly. You can also go in and manually set the velocity change that you want from start to end; a dynamic change of 1 is practically inaudible, if you want to avoid fiddling with invisible markings.
Furthermore, (de)crescendos and the dynamics that precede/follow must be anchored to the same staff. The reason why the dim. in m. 35 doesn't work is because the p dynamic is anchored to the bottom staff, while the dim. line is anchored to the top. This is also why you get the sudden dynamic spike in m. 36. Anchor it to the top, then use the arrow keys to move it where you want. It'll work properly when you do that. This is also why your gradual crescendo has no effect- the end dynamic (f) is anchored to the bottom staff.
In reply to The crescendo and decrescendo by LuuBluum
Completely missed your comment before filing that issue regarding the dynamics spike. You are completely correct. And thanks for also mentioning that the gradual crescendo is on a different staff than the end (f) dynamic - might have missed that, since right now the gradual crescendo ranges from (f) to (f).
This seems to be an actual bug - whittled the problem to a minimal test case - and created this issue .