After 'sf' dynamic, notes remain loud, despite a subsequent 'p'

• Jun 17, 2021 - 23:36

In a transcription of Beethoven's Moonlight Sonata Movement 3, the first measure begins 'p' with an 'sf' near the end of the 2nd measure. The 3rd measure begins with a 'p' dynamic, but Musescore is ignoring it, and the next few bars are played loudly.

I thought perhaps it was because the first 16th note of the 3rd measure is a rest, but even when I apply the dynamic to the eighth note on the bass clef, it is still ignored. If I add a 'p' to the 2nd 16th note in the 3rd measure, it is followed, but the first bass 8th note is loud.

If I change the 'sf' dynamic to an 'f' the piece plays as expected.

OS: macOS 10.16, Arch.: x86_64, MuseScore version (64-bit): 3.6.2.548020600, revision: 3224f34

Attachment Size
moonlight_Sonata_mov_3.mscz 66.64 KB

Comments

• Jun 17, 2021 - 23:51

Apparently, the sf family of markings are subject to some different interpretations from what some of us were taught. In my education, sf markings were described as "vertical" effects, taking place at a single moment in the score. MuseScore takes the approach that sf refers to a range of notes; horizontal, as it were. As it was explained to me, they cannot be changed as freely as regular dynamic markings. If you place the new dynamic a note or two later, it will function as expected.

I'm sure I have not explained this correctly, but I'm sure someone with a better understanding of the MuseScore implementation of dynamics can clear it up.

[EDIT] Think of it like this... when there are two or more interpretations of a musical symbol or effect, the developers need to choose one to implement. This will likely lead to questions along the way, but usually they choose the interpretation which reflects the most common usage.

Reply

• Jun 18, 2021 - 05:28

BTW I dug into how the code works here - an sf dynamic sets the dynamic to 112 then introduces a -18 point diminuendo, so the next note is played at about 94.
I worked out a way to fix it so that after any "sf"-type marking it restores the previous velocity, so it's something we could fix easily enough for MU4.
I do wonder what fp should do though - it's pretty rare you'd have a f passage with an fp marking on a note then nothing on the next note - I'm not sure how I'd perform it either! If it were piano music and certain if it were a rapid passage I'd assume everything following the fp should be p (given you can't do fp on a single note on a piano!). Probably in general I would return to p - it's only dynamic markings starting with sf and rf I'd return to the previous dynamic for the next note.

Reply

• Jun 18, 2021 - 16:08

That p in the following measure does work, but does not return to the velocity (volume) you want. Those parameter(s) are adjustable. The other postings on this page explain more.

In general, MuseScore -- like all software -- requires many default settings for a decent 'out of the box' experience (especially for any new users), but tons of stuff is accessible to customize to one's preferences.
So not really considered 'workarounds' -- more like 'changing the defaults'.

Reply

• Jun 18, 2021 - 06:54

@Nevin Williams... You wrote:
The 3rd measure begins with a 'p' dynamic, but Musescore is ignoring it...

MuseScore is not ignoring it, but definitely playing it louder than the p at the opening measure.
As mentioned, it plays the louder p at 94 (instead of 49). Not only here, but elsewhere in the score wherever sf, is followed by p.

Simple fix,,,
Click on the sf and, in the Inspector, adjust the 'Change speed' to Fast.

Pro fix...
Keeping the 'Change speed' at Normal, you can use the 'Velocity change' to make the p notes that follow the sf, to play at whatever velocity you prefer.

Reply

• Jun 18, 2021 - 09:41

Can't quite understand why changing the speed to fast would help, but changing the amount of velocity change would. p is 49 by default and sf 112, so you'd need a velocity change of -63 to get back to piano.
But in reality an sf marking in a piano passage probably shouldn't be that much louder anyway, I'd think even just 100 is plenty. Though at least on my system 49 is super quiet, I'd probably go with 60 for piano and 100 for sf with -40 velocity change.
Is there a straightforward way to change the defaults here?

Reply

• Jun 18, 2021 - 15:45

>Can't quite understand why changing the speed to fast would help...<

Changing to 'Fast' makes the sf, behave as "vertical" - see toffle's Jun 17 post above.
This is the 'simple fix' . You can right click one sf , then Select->More...->Same subtype and fix them all. No velocity calculations needed.

>... p is 49 by default and sf 112, so you'd need a velocity change of -63 to get back to piano...<
This is the 'pro fix' -- which involves calculating specific velocity setting(s) and enables returning to a p velocity of your choice.

>Though at least on my system 49 is super quiet...<
You can change the value of the p (or any other dynamic), save it to your palette to replace the default p. You can even have several p's in the same palette and use the tooltip to diffentiate.
See:
https://musescore.org/en/node/322246#comment-1084582

Reply

• Jun 18, 2021 - 21:00

True though for some instruments 60 is probably too loud for p.
And yes I figured out why making the change "fast" solved the original problem, though I'd argue there's definitely unexpected behaviour here. I'm not sure it's really necessary for sf to have an associated diminuendo, it's pretty unusual to see a bare sf on a sustained note for an instrument capable of performing a diminuendo. Though admittedly when playing this particular piece the note after the sf probably needs to be slightly louder than p just to be heard given the sustain pedal usage.

Reply

• Jun 18, 2021 - 15:58

I prefer dynamics to be followed as they appear in the score, and not ignored. A human would not ignore the leading 'p' in the the 3rd measure, and I didn't expect Musescore to ignore it either. Any interpretation of what 'sf' is supposed to do should have been overridden by the subsequent 'p' but it wasn't until that adjustment was made.

Reply

• Jun 18, 2021 - 16:29

Again, MuseScore does not ignore it. Play the score and view the volume meter in the synthesizer. The volume does in fact drop.
You prefer differently from the default setting, and so the good news is that you can have MuseScore interpret it as you see fit. The 'how to' is already explained elsewhere in this forum thread.

(Also, read toffle's post above about interpretation of sf. )

Reply

• Jun 18, 2021 - 17:29

So, I took a screen recording of the synthesizer volume window, playing the two excerpts below. One has a 'p' dynamic, one doesn't. The volume bar does not change significantly between the two versions.

Since the volume remains the essentially the same whether there is a 'p' or not, that meets my definition for the 'p' being ignored. The .gif animation I included was rendered at half-speed; the two screen caps are overlaid on each other. They're synced within a frame or two of each other.

While there are multiple ways to interpret the 'sf' dynamic, I'm pretty sure the 'p' dynamic is unambiguous. The volume does not drop down to 49, as it explicitly should.

As a side note, when I changed the tempo of the excerpt to Lento in the course of collecting this data, the excerpt with the 'p' dynamic played as expected, so there is a tempo dependency as well.

Attachment Size
dynamic.gif 794.99 KB
moonlight_Sonata_mov_3-exerpt-No-p-dynamic-at-all.mscz 14.97 KB
moonlight_Sonata_mov_3-exerpt-with-p-dynamic.mscz 15.03 KB

Reply

• Jun 18, 2021 - 19:13

It might be possible that something else is also going on. I spent more time than it may be worth to mess with this.
I opened your excerpt with dynamics. I added to more two more pianos. I copied the right hand to piano two and the left hand to piano 3. Then I played only piano 3. As expected, playback was not at p level. But when i played back pianos 2 and 3, something odd happened. At least on my system. Piano 3 at measure two played at the same louder volume as it did when played alone. But at measure 4, the volume dropped. This only in piano 3.

I took out the repeat barline. No change. I put a p marking at measure 2. Volume dropped even more at measure 4. This says to me that it may not be a case of MuseScore not playing dynamics, as much as it may be that something is keeping it from playing them.

I think the next step would be to open a blank piano score and enter the notes from scratch. But it's not worth it to me. I don't do transcriptions, and don't think I've ever used sf.
I report it here if anyone is interested.

Reply

• Jun 18, 2021 - 21:02

@Nevin Williams...

This discussion concerns playback, which means how one hears, (or wishes/prefers to hear) the playback -- related, of course, to how one would play it themselves. So we actually infer personal preferences.
Your gif lacks sound, so it is difficult to imagine what's sounding at any given moment, although I think I can "see" the sf being played ;-).
In either of your cases (whether or not the p follows the sf) the volume does drop after the sf, but...
...you want it to drop to 49 - i.e., your 'regular' p value. That's whv the Inspector exits. There you can tweak parameters to play according to what you as an individual prefer.

So...
I have attached a simple test file of a square wave tone subjected to various Inspector tweaks. Feel free to change values and listen to the resultant playback to discover what the various Inspector settings accomplish..
(Also, you are correct about a tempo dependency. The first 2 lines of the test file show it.)
Test file: sf_settings.mscz

Finally,
If you want your sf, to behave consistently (e.g., taking place at a single moment), you can CTRL+Shift drag a 'fast' sf, or a 'zero velocity' sf, to your palette to replace the existing one; or label the tooltip to distinguish it from the existing one. This way any p (or other dynamic) that immediately follows sf will be played the way you want.

Reply

• Jun 18, 2021 - 21:04

That actually makes sense. By default every sf in musescore has an accompanying diminuendo and if that extends past the next dynamic marking it overrides it. Definitely not expected and worth raising as a bug. But further, you shouldn't even need the p after the sf, that's just what it should return to automatically.

Reply

• Jun 18, 2021 - 21:37

Thanks, Dylan. I'll submit a bug later. I don't mind so much that the dynamic doesn't return all the way back to 'p', but that the explicit 'p' afterwards was not being honoured.

As an aside, I first transcribed this piece before the 'sf' and other 'r' and 's' dynamics did anything at all, and had compensated for the playback of the 'sf' parts with accents. I only noticed the brokenness when I went back to export this piece again on a newer version of Musescore.

Reply

• Jun 18, 2021 - 22:43

@Dylan Nicholson1... you wrote:
But further, you shouldn't even need the p after the sf, that's just what it should return to automatically.

If you open my example, on the third line, where the sf velocity change is set to '-40', the final p can be deleted with no ill effect.
So, change that value to -63, and the velocity becomes 49 (112 minus 63) which, as you know, is the starting p velocity. So, the final p after the sf is not needed when the velocity change is specified.

The OP's original (Beethoven) attachment has a sf velocity change set to -18 so it sounded to him like the p was ignored. It wasn't ignored, but played at 112 - 18 = 94 which is far from what the OP was expecting -- namely a velocity of 49 (i.e., the same as in the first measure).

You can open the OP's original and set the velocity change of the first sf from -18 to -63. Then the following p can be deleted. This is the "pro fix" I mentioned earlier.
The OP used the "simple fix" (sf set to 'Fast') -- which needs the following p (velocity 49).

Since playback has vastly improved over the years, you raise valid points about the complexities involved.
For instance:

Though admittedly when playing this particular piece the note after the sf probably needs to be slightly louder than p just to be heard given the sustain pedal usage.

But in reality an sf marking in a piano passage probably shouldn't be that much louder anyway, I'd think even just 100 is plenty.

...49 is super quiet, I'd probably go with 60 for piano.

...for some instruments 60 is probably too loud for p.

Phrases like "probably needs"... "I'd think"... 'I'd probably go with"... and what others write, all point to personal preferences. There are no "defaults" which will suit everyone. So, the Inspector is there to allow people to finesse things.
Depending on the goal, one may have to plod along tweaking this or that to figure stuff out; but if playback is important, it's worth the effort. MuseScore has a lot to offer out of the box, though if one wants its synthesizer to sound like a recording, or to play something the way one would play it, then "default stuff" will invariably need to be changed to one's liking.

Reply

• Jun 18, 2021 - 23:21

Sure, but there's definitely a bug that if the diminuendo that's associated with an sf extends in time past a specific dynamic like p than that dynamic is effectively ignored. There's no good case for that happening that I can see. It's not too hard to fix, probably requires a change to ChangeMap so that addFixed( ) effectively truncates any velocity changes that had been generated from a previous addRamp(). And I'd still say the default for a passage that starts p and has a single note with sf and no other dynamics should definitely be to return to p after the diminuendo (which is actually what I thought the poster was originally asking about).

Reply

• Jun 19, 2021 - 00:49

@Dylan Nichokson1... You wrote:
There's no good case for that happening that I can see.

I see where you're coming from...
I neither know nor understand the logic behind the decision to make the sf affect what comes after it. My thought is that even if I want a p which follows a sf, to play slightly louder than a "normal" p, I could change the velocity of the p itself. Why does the sf even "look ahead" and affect what follows? (Earlier there was mention of a more "horizontal" interpretation of sfz.)

Also, the exact rationale behind the 'Slow', 'Normal', and 'Fast' choices available in the Inspector eludes me.
There must be valid reason(s) for this as someone intentionally created those parameters.
I just work through them, as when I grappled with the OP's playback issue: Moonlight sforzando.mscz

Reply

Do you still have an unanswered question? Please log in first to post your question.