Customize Dynamics

• Mar 12, 2023 - 04:32


I just started using MuseScore 4 and I discovered issues with dynamics. For some instruments this is more noticeable than for others. And differences are not similar/consistent across instruments. I'm using Muse Sounds btw. So dynamic levels, especially the mid-range ones, meaning mp, mf, and forte, are too far apart.

For example, take a look at the Muse Sound Grand Piano. Jumps abruptly from mp to mf. Now take a look at Muse Sound Trombone. Piano is barely audible, so is mp, mf is finally a gentle p.. maybe mp, FORTE.. is a huge jump up so it sounds crazy metallic and exaggeratedly loud, ff is pretty much the same as f. The Violins 1 also jump a bit from mp to mf, but overall they're more even. A few instruments like the Doublebasses have extremely soft pianos.. so on.

This creates a problem with both balancing sound, in general, and with Crescendos or Diminuendos that sound surprisingly loud or soft, as a result of having to cresc or dim too fast at times and too unevenly. I believe the jump from one dynamic level to the next should be gentle and not that noticeable, so that a mf < f, or mf > mp is a gentle, barely noticeable effect, and a p < f sounds pleasant. Instead, the difference between mp - mf or mf - forte atm is, for many instruments, so large that it sounds jarring. Across different articulations the issue seems to be consistent with the default, natural sound. So the violins in arco legato, arco staccato, and pizz for example, sound equally uneven when moving from mp to mf.

I think the Crescendo between any 2 levels should be gradual, linear, so that it's reliable. But ultimately, no matter how you set these up, it would be best to have manual control over them.

So my main question would be: could I set myself [in some text file?] the intensity of every dynamic level from ppp to fff? I would want to make piano more noticeable and forte less so, so there is a reason to use dynamics from ppp to p and from f to fff, and so the main range, from p to f, is gentle and even enough. But then some other composers might need different ranges, like a Mozart might only need a p-to-f range, while a Ligeti is more of a pppppp-to-ffffff range.

And as a whishlist, I could add: would it be possible to customize the dynamic curve itself, for a Crescendo or a Diminuendo? Maybe I want it more like a spline.. or more linear? It would be amazing to test various options for it. And I think it would be ideal to be able to customize each Cresc or Dim in the score, individually, if necessary. For playback reasons. Alternatively, customizing the velocity of each track, in a solid, reliable way, would be a more advanced way to handle this issue, but ideally I'd love to never have to touch that and instead, just work visually, like a composer + pen + paper.


After doing some more testing with not just MuseScore but other notation software as well, I have some bits of advice for anyone interested in my bits of advice:

  1. don't do what I tried to do, don't use a notation software for sound output. the results are always problematic, even if you use VST instruments. low quality of sound output, various issues, even in the best notation apps.

  2. if you really want to output sound, don't use MuseScore, especially for large ensembles. poor performance and even lower quality sound output, compared to other apps. do some comparisons on your own, with a simple piano track, and use high-quality monitor headset or loudspeakers. the issues might not be obvious otherwise, like if you use cheap speakers or uneven, biased headphones. midi input audio-feedback is a problem in MuseScore too, but that's another story.

  3. if you really need to use MuseScore, don't use its built-in instruments, even if at a first glance some seem quite lovely. if you dig deeper, there are too many uneven and/or incorrect samples. use high quality VST instruments, this way you get even dynamics and a bit better sound. the final sound output is still problematic, but it could be ok for a preview.



I'm having the same issue. It's a tough dilema - do I want to write what I want to hear or what the players should play. Of course I prefer the 2nd option but then I can only hear the piece when played live and sadly that isn't a frequent enough event.

I'll join you in waiting for an answer :(

In reply to by jonathankraka

With MuseScore, the answer is pretty clear to me atm. Ignore sound, write a score.

If you want sound output, you can use MuseScore for a rough preview, ideally with VST because the built-in soundfont is often buggy and dynamically-uneven, but don't get too attached to this sound rendition, don't base your orchestration decisions on MuseScore sound feedback. Good orchestration and a good musical performance needs subtlety and precise control over details. You can't get this is any notation software.

Better, use a DAW for sound. If you want a high quality rendition, input with a piano keyboard, into a DAW, track by track, and then polish the piano roll positioning, durations, and dynamics.

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