Muse Piano accents are inconsistent

[wizofaus]

Issue type

Muse Sounds bug

Bug description

When using Muse Sounds, the various methods of specifying emphasis (accent/marcato articulations, sf/fz/sfz dynamic markings etc.) cause inconsistent results.

Steps to reproduce

1. Create a new solo piano score using Muse Sounds
2. Add a whole bunch of notes (ideally over 30) at the same pitch
3. Add various accents/marcato articulations, dynamic "emphasis" type markings to various notes
4. Duplicate the whole passage and add a regular f dynamic to the second copy.
5. Play it back and and observe that some accents are very subtle (even a "marcato" hat), and others make no difference at all or even make the sound quieter! (e.g. fz when the underlying dynamic is f)

My general expectation:

• Regular "accent" articulation mark could be something between the current level of emphasis and what you get with sfz when the underlying dynamic is mf. Regular marcato articulation mark (^) should be pretty much what you get using sfz (again when the underlying dynamic is mf).
• Probably no need for any difference between fz/sf/sfz (debatable obviously), at least at mf underlying dynamic. Arguably sf should always mean "at least forte for just this note, but obviously louder than unmarked notes".
• At an f underlying dynamic, fz (or the ^ marcato mark) should give something a bit less than the current sffz).

Screenshots/Screen recordings

https://musescore.com/user/7209246/scores/17311960/s/6liQrz

MuseScore Version

4.3

Regression

I don't know

Operating system

Windows 11

Additional context

No response

[Dima-S-Jr]

See also the commentary, #14736 and #14988.

About the fz. It actually works, but instead of affecting a single note/chord, it affects all subsequent ones. In short, currently fz is played exactly the same as mp and mf (the latter are the same in loudness). (To convince you: enter a series of identical piano notes, place mp, mf and fz everywhere, click play and you will not hear the loudness difference throughout.) It's about MuseScore Basic. In Muse Sounds, everything is almost the same, only mp and mf differ slightly in loudness, and fz is something in between mp and mf.

Probably no need for any difference between fz/sf/sfz (debatable obviously), at least at mf underlying dynamic.

I do not know if it is appropriate for me to talk about this, but some sources cite this table as an indication of the interpretation of force/intensity of attack. Of course, it is controversial and useless to prove the correctness of this table, but it would be nice to use this interpretation, as an option, in MuseScore Studio.

[wizofaus]

I do not know if it is appropriate for me to talk about this, but some sources cite [this table]

Well, yes, I wouldn't necessarily completely agree with that either, but it would be better than what Muse Piano has now.

fz is absolutely not an indication to change to a new dynamic at any rate, and you're right, that's how MuseScore seems to interpret it, even with MS Basic. In fact it seems to treat it exactly the same as mf which is very odd.

[DaddyLudwig]

I've noticed that sf at ff also makes it quieter. It's like it treats sf as f for a single note regardless and so at prewritten f, there's no difference between the sf and f and at ff it acts as a reverse accent! So sf only really acts like sf at mf or quieter. As someone who sees sf in ff passages on a regular basis, this is not what I would expect to happen.

2024-05-25-20-35-28.mp4

[wizofaus]

Related: #14988, #14736

[bkunda]

@sampleeditor @matthewreadbass @RomanPudashkin looks like we might need to review which of our accents are triggering what samples (and in what way) for Muse Sounds.
It might also be worth reviewing other libraries to see whether they also behave in a similar way.

For me, the worst is applying an accent but hearing no change in playback. (The precise degree of volume/velocity alteration is obviously somewhat more subjective). Even a uniform change in volume/velocity for all accent marks could be a good starting point.