Accents way too aggressive in new update?

• Aug 11, 2020 - 03:57

Subject sums it up but I feel like the accents in this update are really exaggerated (staccatos are really chopped too) and forte piano dynamic changes don't take effect until a tied note

Comments

• Aug 11, 2020 - 14:18

I've had the exact same problem with my projects and it's starting to get quite irritating. It is quite inconsistent in when it occurs (e.g. accent plays incorrectly on the trumpet in score (a) but not in score (b), and may play incorrectly on the horns or strings in both). There was nothing mentioned in the release notes for 3.5 about changes to the way that velocities/dynamics were treated, so I'm not sure why it has started to occur. Any ideas why it has happened?

Reply

• Aug 11, 2020 - 15:51

Long story for a short answer:

While using 3.4.2 I started getting notifications when I started MuseScore that there is an update available for extensions. Since I don't use MDL I ignored this and wished it would go away. I upgraded to 3.5 when it came out and the notification persisted so I finally looked at the extensions and figured out how to read the name of the "extension" I have and discovers it was the sound font. The only thing is, I don't remember if I noticed the louder accents before or after I upgraded and the sound font update happened some time before the release of 3.5, I never used the alpha, beta or RC.

Reply

• Aug 11, 2020 - 15:57

Both Soundfonts got updated, the HQ one for 3.4.2 too

Changelog (nothing about accents):

0.2

  • Restored missing hi-hat and ride cymbal from "128:056 Marching Snare".
  • Fixed bad loop in "Cello G4" sample in "000:043 Contrabass" and "017:043 Contrabass Expr.".
  • Altered "000:081 Saw Lead" and "017:081 Saw Lead Expr." to reduce detuning effect and improve the sound. Users who prefer the old sound can find it at "020:081 Detuned Saw" and "021:081 Detuned Saw Expr."

0.1.9

  • Worked around MuseScore's broken volume envelope delay phase by removing the use of the delay phase from the following instruments:
    • "000:016 Drawbar Organ"
    • "000:122 Sea Shore"
    • "008:014 Church Bell" - this is the only instrument that sounds significantly different without the delay phase
    • "008:016 Detuned Organ 1"
    • "008:017 Detuned Organ 2"
    • "008:030 Feedback Guitar"
    • "017:016 Drawbar Organ Expr."
    • "018:016 Detuned Org. 1 Expr."
    • "018:017 Detuned Org. 2 Expr."
  • "008:031 Guitar Feedback"
    • Fixed click on note release.
  • "000:022 Harmonica", "017:022 Harmonica Expr."
  • "000:052 Choir Aahs", "017:052 Choir Aahs Expr."
    • Filtered out a high frequency noise present near the start of the "Ahh Choir F#4" sample. Thanks to forum user HuBandiT for the sample edit. Fixes issue #299366.
  • "000:091 Space Voice", "017:091 Space Voice Expr."
    • Fixed panning on low notes. Fixes issue #232886.
    • Removed unnecessary use of duplicated sample zones--two copies of the same sample were playing simultaneously, panned hard left and right respectively. I presume this was done by the FluidR3 author to create a louder instrument, but this approach is unnecessary and chews up additional synth voices. I have removed the extra samples without any loss in instrument volume.

0.1.8

  • Tweaked the use of the modulation envelope on the following presets and their expressive variants:
    • "000:056 Trumpet"
    • "000:057 Trombone"
    • "000:059 Harmon Mute Trumpet"
    • "000:060 French Horns"
    • "000:071 Clarinet"
    • "008:063 Synth Brass 4"
  • Fixed the pitch of the sample at Ab2-A2 in "000:017 Percussive Organ", "008:017 Detuned Organ 2" and their expressive variants.
  • All instruments have been updated to cancel the default SoundFont 2.04 "velocity-to-filter cutoff" modulator. Previously, only the SoundFont 2.01 version of this modulator was being canceled, but now both versions should be canceled when using synths that support modulators. This change only impacts the use of the SoundFont outside of MuseScore.
  • Added missing velocity override modulators in "000:102 Echo Drops" and its expressive variant.

Reply

• Aug 11, 2020 - 18:46

In order to investigate, we would need someone to attach a score in which they can hear a difference between 3.4.2 and 3.5, and tell us exactly which note or notes you hear the problem with

Reply

• Aug 11, 2020 - 20:12

I thought I was just going crazy, but I had noticed this too, though I hadn't given it much thought until now. But, I also find it interesting that I noticed this on the General User GS 1.471 soundfont created by S. Christian Collins (which hasn't been updated in quite a while).

Specifically, I seem to notice this the most in wind instruments on longer-held notes. I've attached a MuseScore file containing 4 bars of trumpet solo. Every note has an accent, but I hear a clear difference in the handling of the accents from 3.4.2 to 3.5.

In 3.4.2, I hear a very clear, punchy attack that quickly goes back to the written dynamic (which is what is expected). However, in 3.5, I don't hear the note back off after the accent. It seems to just stay at the velocity of the accent. (Also just noticed that it doesn't "back off" until another tied note, ex. m. 3-4).

So, I think the problem isn't that the accents are too aggressive, rather that the notes aren't backing off anymore (though I have no idea why this would have changed).

Attachment Size
MuseScore_accent_example.mscz 5.38 KB

Reply

• Aug 11, 2020 - 20:46

Accents do not have any special code to make them attack then back off. All they do is increase the "velocity" of the note. If the soundfont responds to velocity by making the note louder the whole time - and most do - then that is what you get. I don't think General User has anything special to do anything different from this, so I'd be surprised anything different ever happened. But hopefully someone who knows more about that specific soundfont can comment.

EDIT: that said, the default MuseScore General soundfont is supposed to respond to velocity in a different way than that, in order to support single note dynamics, and I can confirm a difference in sound here with that soundfont. Not sure if it's due to a difference in the soundfont or in MuseScore. The tied note is most obviously problematic, there is a sudden decrease in volume on the tie.

Reply

• Aug 11, 2020 - 21:37

Disregard anyone saying that there hasn't been playback changes. I refactored single-note dynamics for 3.5, but hadn't heard of any negative side effects - until now. There were a few minor changes made to the actual playback, but nothing that should have noticeably changed the sound.

What I really need in order to work out what's going wrong is a score or two where you're noticing these problems - preferably minimal examples that reproduce the issue.

Reply

• Aug 12, 2020 - 14:48

From what I've seen, the fortepianos don't play properly only if an accent has also been placed on them. In other words, without the accent the fp plays perfectly normally. As others have mentioned it doesn't back off to piano until the tied note or next note if an accent is attached, as seen example score provided earlier. I assumed the OP was saying that when the fp was accented, it was not changing velocity until a tie (I may have misinterpreted them though) as opposed this problem being consistent regardless of whether it was accented or not.

Reply

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