Musescore 3 accent is MUCH louder, wish this could be an adjustable preference

• Jun 11, 2019 - 00:49

The accent articulation on Musescore 2 was more subtle (default piano sound), and worked perfectly for my music. With 3, the accented note is twice as loud, and I can't use it. There is no volume/intensity adjustment in the accent inspector, can individual note volumes be fine-tuned in the note inspector?


Comments

In reply to by Marc Sabatella

Accents are apparently giving a velocity adjustment of ~32. This is absolutely extreme and does not make sense in real music. 2.x seems to have adjusted velocity by more like ~12.

This is a bug. I don't have a 3.0 version installed to test now, but it's possible this was introduced in the single note dynamics patchset or something else more recent than 3.0.

People shouldn't have to go in and manually override every note's velocity to compensate for this just to get something remotely reasonable.

In reply to by Belteshazzar_

If the velocity setting is offset, the value you write there is calculated relative to the existing value. (Percent)
For example, if the existing value is 64, the value -12 means:
((64/100) * 12) - 64 = 56.32
That is, twelve percent of 64 is reduced from the value of 64.
In this case the value of -50 will result in 32

(Perhaps this division would have been more meaningful if it was x/127)

In reply to by Adalbi Atskanov

Treatment of these is subjective - in some styles of music the marcato is not intended to be louder than sforzato, just shorter. I agree on certain instruments in certain soundfonts at certain dynamic levels, the default accent is too loud. Unfortunately it's impossible to find one value that works well for all instruments, all soundfonts, and all dynamics. Hopefully in a future version more customization will be possible, but for now, there is the Inspector to set velocity and that seems to work.

In reply to by Marc Sabatella

Sure it works, but annoyingly unsatisfactory when you have tons of accents in your score and then have to select them all so they don't play and then select all those accented notes and adjust their volume.
According to Wikipedia "The marcato is essentially a louder version of the regular accent > (an open horizontal wedge)."
I study some jazz book and just tried to enter some score snippets with lots of accents and marcatos that supposed to show what makes jazz sounds like jazz. With the wrong current Musescore accent and marcato volumes it doesn't work at all.

In reply to by Adalbi Atskanov

As I said, different styles throughout musical history interpret the same markings differently. So yes, someday we will hopefully provide more controls over this. But again, the way MuseScore implements marcato is not wrong, it's correct according to how half the world interprets it. Just not the half that wrote that Wikipedia article. Both are possible correct interpretations in the real world. And it's especially true in jazz - marcato virtually always means shorter, not louder, than sforzato in the jazz world.

In reply to by Marc Sabatella

I don't know where you get your information, as you didn't cite anything, so it sounds unfounded.
On the other hand here is the quote from "Music Notation (Berklee Press)":
"There are two types of accent marks: the forzato accent and the sforzando accent. The forzato accent represents the heavier, or stronger, of the two types of accents. Neither of these accents alter the durational value of the note or voicing they attend."
Berklee is well known for the study of jazz and this quote clearly says that there is something wrong about Musescore accent and marcato.
I can give more quotes if you are still in denial.

In reply to by Adalbi Atskanov

I don't think there is any denial going on. I'm not sure you are understanding what Marc is trying to say. Different fonts handle these markings different ways. MuseScore isn't responsible for that. For example, I have a solo violin font that when it sees an accent in the software, changes to a louder more forceful sound that is way too much. As a result, I can't use that font in certain situations. However, the default solo violin adds a more subtle volume that seems more correct. The marcato isn't louder but more separated. MuseScore doesn't always play well with other fonts. But that isn't necessarily MuseScore's fault. Not all of them are made to be used by MS.
It depends on what you are after. Do you mostly want notation or playback. I often have two scores. One is notation I might give real players. The other is to get playback the way I want.

In reply to by bobjp

I'm not sure you understand what our exchange is about. It is about relative loudness and duration of notes with accent and marcato in music notation theory. It has nothing to do with sound fonts. The generic font (and majority of others) has samples of the same volume (or maybe two levels of volume) for all notes. So, a player would play the same sample for a regular note and for the same note with accent or marcato (because it has only one sample for the note). All the player can do is just applying different loudness (and different attack if it can). Musescore plays a note with accent louder than with marcato and that is wrong according to the theory.

In reply to by Marc Sabatella

But FWIW, I also encourage you to continue reading that Wikipedia article you cite. The very next paragraphs go on to make the exact same point as me: that the interpretation can vary depending on context. And in particular, as that article says, "In jazz big-band scores, the marcato symbol usually indicates a note is to be shortened to approximately ​2⁄3 its normal duration, and given a moderate accent.".

Again, this is almost universally observed among professional jazz musicians.

In reply to by Marc Sabatella

If you despise Wikipedia writers and ignore Berklee Press, here is another quote from 'Essential Dictionary of Music Notation': "When the marcato is placed over or under a note or chord, the note or chord is to be played with even more attack, and more marked, than an accent.
I didn't argue against marcato changing note duration, but that notes with marcato should sound louder than notes with accent.
As to your mentioning changing duration of notes with marcato, here is another quote from 'Jazz Theory Resources': "Lines that end on short notes on the upbeat should are usually played short and accented and will often be notated with a “^” symbol."
So, that is just a special case.

In reply to by Adalbi Atskanov

Let's look at it with some logic:
The audible-length of the note is also slightly shortened, as any accent (accent, marcato, etc) must have a quick fade at the end.
On some instruments (key played, plucked, strummed), there is no possibility to control the volume after the note-playing: this effect is achieved by slightly shortening the normal note-length.

In reply to by Adalbi Atskanov

I don't despite Wikipedia authors at all. And as I observed, they tell you the exact same thing I am telling you - in the jazz world, marcato means shorter, not louder, than a standard accent. It's not a special case, it's how the symbol is interpreted in all cases. The fact that the one book you mention happens to only mention that case doesn't mean it is played differently in other cases.

I am not sure why you seem so intent on disbelieving me, when even the source you cite backs me on this. All I can do is repeat to you: based on decades of professional experience, it really is this way, I absolutely promise you.

It's also absolutely true that in other genres, marcato can mean louder. Such is life - same symbol means different things in different musical contexts. MuseScore's playback is based on the genre in which the symbol is used more commonly. It's a relatively rare marking in classical marking, but pervasive in jazz. So the jazz interpretation won out.

In reply to by Marc Sabatella

Not that I don't trust your experience just like that. Simply your words totally contradict words of those who write books. For example, you say "marcato means shorter".
In Dictionary of Music by Hugo Riemann you can find that "Marcato (Ital.), accentuated".
In Harvard Dictionary of Music by Willi Apel there is "Marcando, marcato [It.]. Marked, stressed".
I don't understand how "stressed" or "accentuated" can mean "shorter".

In reply to by Jojo-Schmitz

I know your team can be quite intent protecting your bugs, I guess this bug was a requirement. I would remove 'marcato' from usage, as it seems confusing for you. In books recommended for university students of music such as 'Behind Bars' they don't use this term. Marcato and accent marks are called 'strong accent' and 'standard accent' respectively. In 'Music Notation in the Twentieth Century' they are called strong accent' and 'normal accent'. So, I think we identified three bugs here now:
1. Notes with standard accent marks are inappropriately loud.
2. Notes with standard accent mark are louder than notes with strong accent marks
3. Notes with strong accent marks are inappropriately shorter than should be.

In reply to by Adalbi Atskanov

The first is arguably a bug, I agree that for some instruments at some dynamic levels with some soundfonts, the accent is too loud. Unfortunately we don't have a way to make it different for different soundfonts or different dynamic levels. So I'd be more inclined to say, the soundfont should probably be tweaked so that the amount of extra velocity employed doesn't come across to strong for any sound at any dynamic. Probably very difficult to implement, but if you're volunteering your expertise here, that's a good place to start. Or just try a different soundfont.

As for the rest, once again: these are not bugs, they are valid stylistic choices in keeping with how these symbols are interpreted in jazz and many other styles of music, even if it's not the standard used in classical music.

In reply to by Marc Sabatella

I think your explanation is not valid and misleading, because (I've already explained this above, but you apparently ignored that, if you repeat the same arguments) if you take , for example, default MuseScore_General.sf3 font, you can see that for the standard drum set for the kick 2 it has only one sample . And if you selected it, no matter what articulation mark you put on your kick notes the SAME sample will be played for those notes. Now, I attached an audio file with the kick hits, and you can test that accented hit is twice louder than regular hit, and 1.5 louder than hit with marcato. This is all wrong, and it produced not by sound font as you say, but by your synth (Fuid I guess). So, your code have total control over the volume and it is easy to implement, just change some default numbers in you code.

Attachment Size
KickTest.zip 130.04 KB

In reply to by Adalbi Atskanov

My explanation is completely 100% correct, and any other professional jazz musician would be able to explain to you as well. I have no idea why you are choosing to ignore or disbelieve what I am explaining, but I cannot explain it any more clearly. Feel free to re-read my explanations if you are still confused about why marcato in MsueScore plays back using the standard jazz interpretation rather than the standard classical interpretation.

In reply to by Adalbi Atskanov

Having graduated from a school of music, I find the defence of books that may or may not be definitive, to miss the point. In the real world, when a player is faced with a piece of music, almost everything he sees is up for interpretation. Exactly how fast is Allegro? How soft is that piano marking. How much do I slow down that ritardando? Then there is that pesky accent. What is appropriate? The books MIGHT be a starting point.

As to MuseScore:
1. This is not true in my experience. I have a series of accent marks over notes for the entire orchestra in one place. Volume is only slightly louder. This is with the default font. Where I have seen a too loud accent, it has been a problem with the font. Not MuseSore.

  1. True, but only because (see 3.)

  2. Are you talking about marcato? You can't just stop using a word based on how you interpret some book. That term is not going away. And it is clear that there is disagreement on a definition. This is unusual how?

To be clear, there just aren't that many hard and fast rules in music. You can say that a half not gets two beats. But in practice, that may or may not be true. There are certainly accepted practices. But those change, sometimes.

In reply to by bobjp

  1. As I said above and you can check it, an accented kick sounds more than twice louder than regular one. I wouldn't call it "slightly louder". I'd call it much too loud.
  2. I'm sure majority of Musescore users would be music students. So, if they study notation using books that don't use term 'marcato' they will be confused about it, as all Musescore team is. So, it can easily and should be put aside.

In reply to by Adalbi Atskanov

  1. I'm not sure your wav file helps your case. "Sounds twice as loud" is subjective. And will vary from person to person depending on many things. Are we talking headphones or laptop speakers? Monitor speakers or earbuds? To me, nothing in your wav fie sounds "twice as loud." Perhaps you could post some graphs. I'm posting one of your file. Maybe it will help.
    kick.png
  2. I have no idea who the majority of users are. Nor does it make much difference.

BTW, I am not a member of the "MuseScore team". I'm just trying to follow your observation. As I said, I have an orchestra piece where I have accents on four consecutive notes in every instrument. Twice as loud would have split my head. Instead I heard what I feel to be an accurate accent.

Why stop at marcato? There is no simple answer to what staccato means. Or most any term. Why? Because, as has been pointed out, the meaning of terms depends on the style of music they are applied to. Tempo and volume apply also.

In reply to by bobjp

Thank you for posting the image. I should point that the first hit is regular, the second is accented and the third is with marcato. As you can see the accented hit is at 1.0 and the regular is at about 0.47. So, you could easily calculate that accented hit is more than twice louder than regular. So your image says it all, no additional graphs required. But you can check it all by yourself if you so inclined. And yes you made me laugh thinking that I'd measure the signal level with the help of headphones. But it is not actually unreasonable, because doubling of sound intensity corresponds to the level change of +3 dB. And If I used headphones as you suggested then sensing loudness doubled would correspond to the level change of +10 dB.
+3 dB wouldn't split any head. But for me, as for others, who posted on this thread, it is inappropriately loud and makes some types of music sounds ridicules. That is why I found this thread after a long struggle with these accents in my scores. I agree it can be called subjective, as some people has sharp hearing and some lowered. So, then I see another bug here:
6. Accent and marcato marks have subjectively incorrect default volume values and Inspector for those elements does not allow to set the default values for a current score and globally.

In reply to by Adalbi Atskanov

I made no such suggestion to measure anything with any device. I meant that the same sound will sound different depending on what device you are listening to. This is well known in the recording industry.
So, here are a few more wav. files from the default font. Normal, accent, marcato, normal.

piano acc.png piano
trumpet accent.png trumpet
flute acc.png flute
violin acc.png violin

I see that the piano and trumpet files show a volume difference. Yet, the flute and violin do not. What could be the reason? It seems to me that percussion, piano, and to a lessor extent, trumpet, have a hard start to their sound. You have noted that only some instruments are affected. I suspect this is why.

Your suggestion: " So, your code have total control over the volume and it is easy to implement, just change some default numbers in you code." I can't see this being at all possible. What would a lower number do to the violin sound? Make the accent softer than the note before it? How could the code possibly know when to accent and how much, on a universal level. Unless it was locked down to a particular font that is programed into the code. This is not the purpose of any notation software. If it was as simple as changing some default numbers, don't you think it would have been done long ago?

The problem with any notation software is that by its nature, and for the purpose of playback, you can't always write what you know works. To some extent you have to, what is commonly refereed to as, writing to the sounds.
I agree that there could be better ways to implement some things. It's getting there.

In reply to by bobjp

When you start a playback you can move a volume slider up and down and you can hear volume changing. Right? The same thing happens when MS player plays your score. It parses the score and when it encounters accent mark, simply speaking, it moves its internal slider up and down. As you can see on your images they implemented shortening for notes with marcato, but apparently they didn't implement volume adjustment for accented notes, though there is plenty of room for that.
So, congratulations, you found another bug:
7. Inconsistent implementation of volume adjustment for standard and strong accents for different instruments.
Honestly, I could go into their code to waste several weeks to prove my point, but the problem is like this. I used to know one old guy, who now already left this world. When some officials would visit him trying to make him pay his bills and taxes, he would pretend to be deaf or even an idiot, so, they would leave him alone. The same tactic started to use QA people here. I still see bugs bugging me that I reported long time ago, so why bother?

In reply to by Adalbi Atskanov

Interestingly, you already mentioned the inconsistent implementation of accents. And you misunderstand what I am saying.

You believe that all that is required to make an accent work is a simple adjustment of default numbers. My images show there must be way more to it than that. By your reasoning, MuseScore doesn't care what instrument is playing. It should just see an accent and apply a consistent formula to it so that all accents (of whatever type) sound the same. How do you know this is not the case? Look again at the file for the kick drum. each note is a separate entity. Then look at the flute. The notes are connected. What standard does the software use to base its decision on how much volume to give the accent? If there was such a thing as a default number for accents, don't you think think it would applied to every accent. Accent=default boost.
The problem with recorded files is that they are not real players. Files don't blend together the same way real players do. Isn't it possible that when the software sees an accent on a continuous sound, it doesn't have a base to start from. To me, accent implies attack. The start of a note. Where is the start of the flute notes? Wouldn't a sudden burst of sound in the flute just sound stupid?

In reply to by bobjp

For sure they apply some sort of ADSR envelope to the sound, based on some algorithm they devised. I just tried to simplify things for you for better understanding of the fact that they have control over this stuff. They were able to make marcato note shorter, but I think it's harder to implement than the accent, as it requires multiple parsing. That is why they protect it so hard - it was a requirement (though incorrect) and was not easy to implement. I think, they screwed up something in v.3, because if you check Musescore v.2, as I did for violin from MuseScore_General.sf3 you will see that it's different from your tests and has appropriately loud accent (though still louder than marcato). And my point is this – if they screwed up something they should fix it.

In reply to by Adalbi Atskanov

For the record, there is nothing the slightest bit difficult the implementation of amrcato playback, nor does it involve anything so arcane as ADSR envelopes. It's a simply matter of deciding on how much shorter and how much louder we wanted to the note. With input from a number of professional jazz musicians, we settled on what we settled on.=, and all professional jazz musicians I know of seem happy with the choice. it's true that those expecting classical semantics are surprised, but as I've mentioned, it's much less common in that world, so it honestly doesn't come up very often. Whereas had we made the mistake of implementing marcato playback using classical semantics, jazz musicians would be complaining almost daily, because it's such an integral part of that music.

The reason we "protect" this decision is that time has absolutely proven this to have been the right decision based on the minuscule number of complaints we get from classical musicians. But, once again, it's completely reasonable to hope that someday both options will be available. For now, though, if you prefer the classical interpretation you can simply set that up manually.

Since you say you are doing this in an effort to learn jazz, though, best to simply leave it alone - ask any professional jazz musician you know and they will tell you the same.

In reply to by Marc Sabatella

You previously said that "in the jazz world, marcato means shorter, not louder, than a standard accent", and you liked some statement in "Marcato" article in Wikipedia, which doesn’t have a reference to prove it, and cannot be verified. I actually found the paper called "Interpretation of Jazz Band Literature", from where the statement probably originated.
First, in article examples "accents are added to several notes that need emphasis." (No problem here). Then it goes:
"Staccato and marcato (cap) should be found only on notes with a duration of one beat or less and generally suggest that the note is to be played less than its full written value. The staccato note should be given half the written value and should receive no added emphasis.
The marcato on the other hand is to be given added emphasis, and in most writing, suggests slightly longer duration, approximately two-thirds of a beat which is variable depending on tempo."
So, accents used for "emphasis" and marcato for "added emphasis".
Therefore, we still have the bug #2 (Notes with standard accent mark are louder than notes with strong accent marks.)
And we got another bug here:
7. Marcato should only be applied and played on notes with duration of one beat or less. Currently it played on any note.

In reply to by Adalbi Atskanov

I don't need a reference to prove what decades of experience tell me is true. Once again, ask any professional jazz musician you know and they will tell you the exact same thing. Marcato is shorter, not louder, than a standard accent. There is not even the slightest tiniest bit of debate about this among any professional jazz musician I have ever met. i have played in, recorded with, and directed countless professional jazz big bands and other ensembles, and every single marcato note ever played any any musician in any of those ensembles was played short, I can absolutely assure you. It's not about what random articles you find on te Internet say - it's what actual professional jazz musicians like myself actually do.

So once again, it is not a bug that MuseScore chooses to use the standard jazz intepreretation rather than the standard classical interpretation. It's a valid and conscious choice that time has proven was the right one, if we can only have it one way. but better will be if we someday give the user both options to select between. So I completely support any feature request for this.

Whether or not you also happen to believe any given arranger should follow the subjective advice you read about whether or not to add maracto to longer notes - I happen to agree marcato on half notes are not a good idea normally - - it certainly doesn't make it a bug if MuseScore chooses to honor the arranger's request if the arranger chooses to make it. That is, I agree an arranger should probably not put a marcato on a half note. Just as I would agree with someone who says an arranger should not put a minor ninth interval between two saxophones in a drop 2 voicing, or should not expect a bass trombone player to be able to play notes far above the bass clef staff. But if the arranger chooses to do these things, MuseScore should absolutely respect those choices and play what is written, not what some arranging textbook says they should have written instead.

In reply to by Marc Sabatella

The subject revolves around something else, but the important thing is: Isn't the accent (>) stronger than it should be?

Assuming an accent will take the value of the next-higher-dynamic, the values ​​are expected to be as follows:

ppp (16) + ">" = pp (33) 200%
pp (33) + ">" = p (49) 148%
p (49) + ">" = mp (64) 130%
mp (64) + ">" = mf (80) 125%
mf (80) + ">" = f (96) 120%
f (96) + ">" = ff (112) 116%
ff (112) + ">" = fff (126) 112%

Of course, since the Musescore software uses a single average value for this, a mean value of 120% is reasonable. // assuming mf is the average default

Let's open a blank score and write a note. Let's add an accent (>) and an mf dynamic to this. If we export this score as "MIDI-File", we see that the velocity value written there is 120. We can also calculate that the rate of this is 150%. // 80 * 1.5 (150%) = 120
However, the velocity value should have been 96. // 80 * 1.2 (120%) = 96

It is obvious that the 150% value is higher than required for "accent".
However, if we look at the average in the table above, it is obvious that the 120% value (also used for marcato) is more reasonable.

This is because the value of "sforzato" is defined as 150 (%) in "instruments.xml" and this definition is used for "accent". Whereas "marcato" is defined here as a value of only 120 (%). If both were to have the same dynamic, they would both be expected to be of the same value.

And most likely these definitions apply to template files (and "My fist score" by default) as well. Therefore it cannot be solved by just modifying "Instrument.xml".

In reply to by Ziya Mete Demircan

It's great that you mentioned this instruments.xml. I suspected existence of something like that somewhere there, but didn't realize where to find it. And these guys kept mum about it. (Probably they didn't know about it too). Anyway, it looks like I was correct when I said that "they screwed up something in v.3". In the file for version 2 for sforzato they set 120, but for version 3 they set 150. Somebody is getting deaf!

In reply to by Jojo-Schmitz

So, Mr. Joachim Schmitz, as the major contributor to the file "instruments.xml" development on GitHub, you should know that someone, namely anatoly-os, on Jan 26, 2019 committed changes called "add double articulations as separate type" to the file. After those changes velocity for sforzato got changed from 120 to 150. So, I wonder what was the reason for that change that caused this entire hubbub on this thread.

In reply to by Adalbi Atskanov

I might be able to have found that out myself, but I'm surely not capable of remembering every change to any file that once I touched myself.

For the record, the commit is a2d52d1

If you what to know why Anatoly change that, you'd have to ask Anatoly, but the addition of those double articulations indeed seem to be the reason, and not an invalid one.

In reply to by Jojo-Schmitz

Jojo, the page "instruments.xml documentation" at musescore.org says that "The document is not really aimed at developers but at MuseScore users familiar with XML and wishing to set up their own instrument definitions." As Ziya Mete Demircan already mentioned that "Therefore it cannot be solved by just modifying "Instrument.xml". I also tried changing velocity in different articulation tags, and it didn’t work. Do you think it is not supposed to work or we need to do something else?

In reply to by Ziya Mete Demircan

That's great. Thanks to you and Jojo for tips. I played with instruments.xml for a while and got better picture. And you were right, when you said "it cannot be solved by just modifying "Instrument.xml".
So, the question is what to do if you are unhappy with how accented notes sound? What if you can hear inconsistency in articulations, say for lead guitar and bass guitar? How to fix this? I see three scenarios:
1. If you create a new score with just a couple of instruments selecting them directly, then MuseScore will copy articulations from instruments.xml to your score. So, by editing the file you get articulations that you like.
2. If you create a new score from a template, for example, 01-Rock_Band.mscx, then MuseScore will copy articulations from that template to your score. So, you can edit this template to get balanced sounds for your lead and bass guitar.
3. What if you created a score with lots of articulations and found that it sounds bad with accents too loud and marcatos too quiet? What if your professor taught you that marcato should not make your notes shorter? What if you have many scores like that? You can change the volume of each note tediously as somebody suggested, getting irritated more and more, but still you cannot cancel shortening your notes by marcato. There is an easier way - you can export your score to uncompressed MuseScore file and edit articulations there and open it in MuseScore again.
Apparently, you can find this reckless velocity setting for accent articulations in many templates, and that is why MuseScore team will not deal with this problem.

In reply to by Adalbi Atskanov

Your claim: Apparently, you can find this reckless velocity setting for accent articulations in many templates, and that is why MuseScore team will not deal with this problem. is just nonsense, esp. the 2nd part, the fact that those setting are used in the templates has got nothing to do with whether "the MuseScore team" will or will not deal withthe issue. Actually here you come in as a team member: fix those templates (and instrument.xml, esp. in the case of non-SND-capable instruments) and submit those changes as a Pull Request.

In reply to by Jojo-Schmitz

The answer to your answer is in your answer. If this is "nonsense", then why to implement it? The "reckless velocity setting" is considered by some here as just fine, no matter what others say. And you propose to take care of this "nonsense" to me, a simple user, who just wanted to get a decent sound to his scores? That is why I said nobody will deal with the problem.

In reply to by Jojo-Schmitz

As Mr. Sabatella remarked "I seriously doubt any of the developers on this thread were unaware of that file". And I can imagine how "they" chuckled seeing how a bunch of blind men tried to figure out solution to the unclear problem. For sure "they" knew what could be done about it, but "they" kept mum, and only woke up to dump a piece of demagogy on confused users.
So, I figured out the problem and solution for myself and shared it, I guess, I can retire from the thread now.

In reply to by Adalbi Atskanov

Please do.
But as you recall, early in this thread it was pointed out that MuseScore is open source. That means that you are more than welcome to modify it as you see fit. Nothing about it is set in stone. There is no "users v developers" mentality here. There is no set development team. No boss. No dutiful minions running around trying to cover for mistakes. I find it interesting that you believe that anyone who disagrees with you is blind and ignorant. Grow up. Your teacher taught you certain things. Great. Did your class talk about how loud forte is? What is the db level? How fast is allegro? Exactly how loud is an accent? Are the opening notes of Beethoven's 5th accented or played as a pickup. There's lots of discussion on that. The fact is that, just like MuseScore, music is not set in stone. For the items above you can probably tell me what each isn't. But you can't say exactly what each is. There are too many variables. There isn't notation software out there that can deal with all of them. Nor do they have to. That's why you have to run your music through a DAW to begin to get good playback.
That's also why it is impossible to claim that one book is the absolute authority on music theory. It's that author's opinion based on their experience. That's music. It's your experience. Everybody gets to decide what music is for themselves. Theory and notation is a way to get musicians on the same page, so to speak. In the end the musician decides what is right. Not the notation. Not the theory. Not the software. Not the composer.

In reply to by bobjp

All you just said is the same crappy demagoguery.
Don't put your distorted thinking in my mouth. I joined this thread because I wanted to find help with what I thought was wrong accent velocity as well as marcato velocity. That's it. You and Mr. Sabatella tried to prove me wrong and dragged me into useless discussion. Though I cited about seven best existing notation references, in response all I've seen was just references to personal experiences. That's not right, but OK. Everyone is free to express their opinion.
When I called users "blind men" I meant myself in the first place, because I didn't know too much about this problem and fumbled around to find any solution. I guess you are unhappy because I said you wasted my time. I still think so, and your reply just confirmed that. Did I get any help from you in several threads? Nothing but frustration.

In reply to by Adalbi Atskanov

Switching to DAW to get a better sound is a dumb idea. This could say somebody very unexperienced with music software. I created several sound fonts with unique instruments. So, what is the difference in sound if I use them in DAW or in Musescore? In DAW it would be Sforzando VST (or something else) in Musescore – Fluid. Which sounds better? I actually think that I get better sound for my fonts in Musescore. That is why I care about how Musescore works.
I also know that I work much faster and more productive using Musescore than DAW. I come to a conclusion, if one knows music notation he'd prefer a music notation program. Otherwise I'd switch long time ago. Of course, I will have to work with DAW if want some special sound FX or have complicated scores and in many other cases. When REAPER will get better notation plugin I'll start thinking. But at this moment I don't see how in DAW I can get better sound with my sound fonts.

In reply to by Adalbi Atskanov

Thank you for allowing me to have an opinion. No one dragged you into anything. They stated their opinion based on actual musical practice. Hopefully authors do the same. All you had to do is not respond any more. Much like most others in this thread.
" I guess you are unhappy because I said you wasted my time. I still think so, and your reply just confirmed that."
You flatter yourself.
BTW, have you finished reading about the Synthesizer?

In reply to by Adalbi Atskanov

I have no idea why you continue to resort to insults.

We didn't mention this file because it's only one of many ways to workaround the "problem", and not one I'd personally recommend in most cases - the Inspector is usually better. The relevance of the isntruments.xml file is for how we might add more control and tweak the defaults in the future, not as something I'd expect a user to mess with.

Mostly, though, I was simply trying to educate you about the jazz interpretation of marcato. That is not a problem to be solved or worked around - it is simply something for you to learn about, and we're here to help with that.

In reply to by Adalbi Atskanov

No one, and I mean no one, has even claimed the settings are perfect. As I have stated many many times, indeed, the sforzato accent is too loud for certain instruments. Marcato is perfect, and that is where most of the discussion has centered, but we all agree that for some instruments at some velocity levels, sforzato is too harsh. No one in the entire history of humanity has ever denied this.

So again, please please stop with the false attacks. If you want to contribute to this discussion, start focusing on solutions - suggestions for algorithms to help sort out which instruments at which velocities should have which percentages applied, so we can improve future versions

Attacks accomplish nothing except turn off the very people who would be actually helping to implement this and make them reluctant to want to. Plus it drains our energy, energy we would need in order to do the word.

Please please please please please just stop with the attacks. Please.

In reply to by Marc Sabatella

I didn't attack anyone. If you feel yourself attacked when I challenge your ideas, it's not my fault, but your insecurity caused that. It amazes me how many contradictions you have in your argumentation, and yet you are not able to see them, and when somebody points at them you get offended. So, it's not me, it's just your inflated ego to blame for your feelings.
I saw your quite old post where you solicited help about instruments.xml, so by now you should know lots about the problem, but did we get help from you in this matter? No. Now you don't recommend using it. Ha.
You said that you made a decision to change all articulations in favor of jazz style. Then there is a question, how come then chamber music got affected, or orchestral music. As I have learned, all music styles have separate template files, why all of them became jazzed up? I understand it's easier to just copy and paste in all them the same articulations, and then to blame users if they are not happy.
For sure, among yourselves you know who made the decision to change sforzato velocity in ALL template files from 120 to 150 and it wouldn’t matter who, but it's funny to watch now how you in turns get hysterical just to shift attention from the fact. People complain about the problem, but you just say:"150 does sound good".
So, don't take it personally.

In reply to by Adalbi Atskanov

I am here willing to help. Please review your own comments, though, and consider whether there is any sort of world in which one would not perceive an attack or insult in statements like "I can imagine how they chuckled seeing how a bunch of blind men tried to figure out solution to the unclear problem. For sure they knew what could be done about it, but they kept mum, and only woke up to dump a piece of demagogy on confused users". I don't think there is any such world. And that's just one example.

This isn't me being insecure, that is you being incredibly and unbelievably rude for no reason at all - we are simply trying to help, and don't serve your attacks.

So I'm done with this thread; I will not look at further response here as it's too hard to follow at this point anyhow. If you wish to start a new one in a respectful tone,. starting with an apology for your unacceptable attitude here, and following up with a clear statement of the problem you are looking for assist with, I'm happy to help.

In reply to by Marc Sabatella

I don't want to be seen as a backbiter here while you already left the thread, but still I want to give a final assessment of your "help". You intentionally tried to misinform users when you said that there was no way to change default values for accents, that it is hard and maybe in some future we will have some control over it. But it became clear that mechanism for changing those values was created long time ago in form of standard and user templates. And you knew about it.

In reply to by Adalbi Atskanov

"I don't want to be seen as a backbiter"
Then why this post?
" You intentionally tried to misinform users"
This is pretty funny and impossible to prove. Both are spoken by someone who always has to be correct and have the last word. I'm surprised Marc stayed in this as long as he did. He took the high road in dealing with you while you resorted to insults. He has better things to do.
When someone askes a question on a forum, the respondents have no idea of the experience level of the person asking the question. It could be easily said that you "intentionally" misled respondents by saying that all that is needed is changing a number in the code. You will disagree and come back with some comment about me. Then I will "intentionally" make some false claim about you. And on and on.
There, I've just taken care of the next three or four posts in this thread. Don't we both have better things to do?

In reply to by Adalbi Atskanov

I seriously doubt any of the developers on this thread were unaware of that file. Please, don't report to ad hominem attacks and insults, it really doesn't lead to intelligent discourse or problem solving.

I suspect the change in the velocity setting has something to do with what is needed to support single note dynamics, as velocity isn't used for them the same way it is for piano. That is why I keep observing the effect is more pronounced for some instruments than others. For trumpet at mf, for example, it's actually quite convincing and pleasing out of the box, because the velocity affects only the initial attack but then immediately backs off. It is mostly the instruments that don't support SND - like piano - where the harsher attack is more problematic.

Probably the definition for the piano specifically - and whatever other specific instruments are found to be problematic - should override the default. Either than or there could be two different settings, one for SND and one for non-SND.

In reply to by Jojo-Schmitz

Interesting! I suppose it's possible this was a case of a change being made during development for 3.1 but accidentally merged to 3.0.3 branch. Probably the history could help sort out whether that is at all plausible.

But in any case, the fact is, 150 does sound good for many SND instruments, probably considerably better than 120 would.

In reply to by Ziya Mete Demircan

Yes, I think we all agree that for at least some sounds and for at least some velocity levels, the sforzato accent is louder than we'd like. Not true for all sounds or all velocities, but definitely for some. There is question about what the best way to address this is, but very possibly this will change more fundamentally for MuseScore 4, so I'm not sure any suggestions based on the current implementation will be applicable.

In reply to by Ziya Mete Demircan

This brings up the question (see my other reply from today at the bottom (currently)) whether the dynamics should be spread differently: instead of an even distribution by absolute numbers, use a distribution where the percent difference between the levels is similar. Having pp double as loud as ppp sounds wrong, but I don’t know the acoustic models used by MuseScore, MIDI, and the synthesiser/soundfont complex here.

In reply to by mirabilos

I’ve experimented a bit with this.

 10 pppp        180% to next level (ppp)
 18 ppp         183.3̅% to next level
 33 pp          133.3̅% etc.
 44 p           127.2̅7̅%
 56 mp          126.8%
 71 mf          125.3%
 89 f           122.4%
109 ff          105.5%
115 fff         104.3%
120 ffff        105% to next level (fffff)
126 fffff       100.8% to max velocity (127)

With this, we can have the accent at 120% and the marcato at 150% (though the default velocity must, most likely, be lowered).

For the Jazz “marcato”, there are two options.

One, use U-0001D180 𝆀 MUSICAL SYMBOL COMBINING MARCATO-STACCATO for it and then just don’t render the dot in MuseJazz.

Two, make it an inspector property, just like trills have Baroque (start from the upper neighbour note) and Modern (start from the note itself) styles that are visually not distinguished.

I think these should make even Marc happy ;)

I’m not sure how much of this we can currently do in the end user role, with a custom workspace with dynamics changed accordingly, custom instruments.xml and possibly an initial dynamic. Might be interesting to try.

I’ve placed the ranges so they still lie within the common dynamic bands used with soundfonts (2/3/4-layer ones) and compared with these other software uses. I’m no sound technician, so I don’t know how well these will work in practice…

In reply to by Adalbi Atskanov

the way you are attempting to force some form of blame onto the developers of musescore is honestly just childish. here you are arguing using supposed references and books, trying to point out "bugs" in musescore. Your persistent arguing in anyone's eyes is (I don't mean to be rude) just childish and (coming from me a teenager) if you have issues with musescore, just go make your own app since (according) it is easy to program it right? FYI you sound like my stuck up conductor who always thinks he is right. It's annoying.

In reply to by Adalbi Atskanov

Once again, I am not saying "marcato means shorter" or that anyone who says otherwise is wrong.

To be 100% clear: I am saying there are two different and equally valid interpretations. One is the one you see when you consult primarily classical references. The other is the one you see in most jazz references and is used by virtually all jazz musicians - that is the one that says to play shorter, like Wikipedia acknowledges.

It's not that one interpretation is right and the other wrong, they are both right, for their own context. MuseScore had to pick one to be the default. No matter which was chosen, those expecting the other would be disappointed. The choice was made to prefer the interpretation used by the musicians who actually use this symbol regularly - it is, again, relatively rare in classical music compared to jazz.

In reply to by Marc Sabatella

If you selected jazz style interpretation for MS notation then you should be consistent. For example, let me cite again one of the best 'Jazz Theory Resources': "A traditional jazz band will play four quarter notes in a row and beats two and four will get slightly more of an accent than beats one and three. Quarter notes will usually be played short regardless of their location in the measure".
So, we get two more bugs here:
4: Off-beats quarter notes played incorrectly because they are not accented.
5. Quarter notes played incorrectly because they are not shortened.

In reply to by Adalbi Atskanov

You are describing here some extreme subtleties in terms of how a jazz player might tend to naturally articulate things. That statement is actually way way way too broad. Yes, there are some types of jazz contexts in which those statements might be somewhat true, but in equally many, it's not true at all. Tempo and style have a ton to do with it also. We don't do that sort of AI processing, although I know people have created "humanizing" plugins.

Anyhow, it's not that we always apply jazz interpretation - most scores aren't jazz, after all. It's just that in the specific case of marcato, given the two conflicting definitions, we did choose the jazz interpretation for exactly the reason I have already stated multiple times. It's a much more common notation in jazz than in classical, so jazz players would be much more likely to notice if the playback seemed "wrong" (based on their personal subjective expectations) than classical players would.

Someday, again, we will likely provide more user control over all of this. But for now, we have made the best compromises we could based on the feedback from thousands of musicians over the years, trying our best to produce results that satisfy the greatest number of users.

In reply to by Marc Sabatella

Here Ye Here Ye ! Context matters in varying music(s) ! Beethoven's published German scores used marcatos where I see in modern English scores staccato . Performers come to their own conclusions and jazz players do the same . Intuition counted for many of Rosalynn Tureck's guesses and time and later
discoveries of Bach's amendations proved her right in some cases . Rules and laws can be besides the point in real life ! You guys are much appreciated !

In reply to by Marc Sabatella

Apparently, there are two drum technique called 'Accent Drum Technique' ("Drummers have the ability to “accent” different parts of rhythms by making certain notes sound louder") and 'Marcato Drum Technique'("A marcato technique is represented by one note being played more loudly or more forcefully than others surrounding it"). Besides, playing a kick drum notes short doesn't make any sense. So, the current implementation of accent and marcato for drums and percussions still has three bugs.

In reply to by Adalbi Atskanov

You have just described the reason marcato - and staccato, for that matter - don't make a lot of sense for jazz drums. Again, not a bug, just a notation that is nonsensical to begin with. Any notation intended to express note duration doesn't apply to drums for obvious reasons. Occasionally you will find them notated anyhow for consistency with how the rest of the band is articulating, but it's actually a bit counterproductive, because then drummers have to mentally ignore some of the markings while honoring others.

In reply to by Belteshazzar_

I just checked 3.2.3 (which doesn’t have configurable offset for accents yet), and the amount of accent is indeed relative to the base dynamic. This is a major WTF.

If I sing/play an accent, I just make it a bit louder and more intensive, but the amount by which doesn’t really differ whether the base note is pp, p, mf or f (ff gets cut off because it’s already so near the maximum). This should be linear, just add a fixed number.

I know I'm way out of my league here, but I've been thinking... and in my case, that is seldom productive. :)

Nonetheless, What I've been thinking of is the MuseJazz style, which I seldom use, so any presumptions I make are likely to be wildly out of place.

I don't know if the MuseJazz style is much more than an interchangeable "skin" which can be overlaid on any given score, or if there is any distinct functionality which belongs to that style alone. What I mean, is I'm wondering whether or not it's possible to assign response to articulations, dynamics, etc. to behave differently under the Jazz style than in the traditional engraving. If this is possible, then the differences in playback could be linked directly to the visual style of the score. In this way, someone who relies on jazz playback can hear those by selecting MuseJazz style, and those who do not, can select the traditional/classical style.

I can think of several objections to this idea, not the least of which being the fact that many people scoring for jazz may not prefer to use the jazz font. Also, I'm sure that for every user who would like to see customized playback as per the visual style, there would be just as many who would rather there be no difference in playback whatsoever.

Feel free to tell me this is a dumb idea. It's Friday, and I'm in a great mood. :)

Coming a few months late to the game here and WOW, a lot of jibber-jabber on this thread that isn't offering a solution to the problem.

The accent on the piano playback is, indeed, ridiculously loud by ANY metric. If anyone happens to see this, can you tell me if there is an easy and logical workaround and/or was this reported as a bug for future attention?

In reply to by tashaffer

As discussed above, indeed, there is a problem specific to piano (the accent works well on most other instruments. The very first response shows how to solve this via the Inspector. There is also discussion of where the issue is in the instruments.xml file that is included with MuseScore. What is lacking, as far as I can see, is a fixed version you could download and install. So someone would need to create one. I recommend starting a new thread with a discussion of that specifically and hopefully someone can help with that.

In reply to by tashaffer

@tashaffer:
One easy workaround is to change the volume factor in the articulations section of the mscx file, (by editing as a mscz). I have done this for nylon string guitar as I also prefer less of an accent. Doing it this way means that you don't have to revisit the inspector method when you add more accented notes.

In reply to by yonah_ag

(I store all my files as .mscx anyway, so I can do git diff and friends on them and edit them.)

The articulations sections I can find in them, v2 and v3, are in the instruments, which come from instruments.xml, and mu͒2 and 3 both have supported a custom instruments.xml file… hmmm… still bothersome to copy and edit every instrument I use and keep them up-to-date… (and, of course, the templates… but I perhaps should create my own anyway)

In reply to by mirabilos

Templates would be easier than maintaining instruments.xml but I was really only thinking about single instrument scores and being able to change the value on a per-score basis. Since I only make solo guitar scores it does make this workaround practical for me.

For me, it sounds better when I change the voice to Acoustic Bright Piano. This can be done by pressing F10 and select voice. It sounds much better and isn't extremely loud.

Bumping this. This is an objective flaw (along with the marcato = loud staccato nonsense) and I'm tired of people trying to excuse these poor decisions. Fix these annoyances.

  1. Marcato does not mean staccato. It should not affect note length. And don't cite jazz rules to me, I know them. They're not standard anywhere else and there should be a separate version in that case.
  2. Marcato is generally interpreted as louder than accent. In MuseScore, this is the opposite.
  3. Accent means stress, emphasize. It doesn't mean hit your instrument with a hammer. In MuseScore, accents are ridiculously loud, giving a 127 velocity at just forte, and pretty close to 127 at mezzo-forte.

In reply to by Arpicembalo

  1. Yes.

  2. Accents (both regular and marcato) usually mean to attack more loudly but then quickly return to the original dynamic. This doesn’t work for all instrumnets, of course. With marcato, this “initial boost” is more (see 2).

(See also Wikipedia for marcato.)

  1. Erk, this is tricky. I think Gould doesn’t help when the only place she describes marcato is in the description of the staccato wedge: “a heavy staccato (between marcato and staccato), also staccatissimo”

She seems to imply that marcato does shorten the length of the notes by quite a bit, by more than staccato.

The only way I can personally bring these things into congruence is that you and WP describe the case for instruments with SND and Gould the case for instruments without SND, which can only emulate the sudden-loud-attack-then-back-to-regular-velocity thing by shortening and loudening the note.


As for the velocity numbers…

… I think, with a mere forte being 96 out of a 0‥127 scale already, we don’t have enough range to maneuver accents at all. The dynamics range is ridiculously skewed in favour of the more silent dynamics. Sure, mp, at 64 sits just on the upper side of the half, but in my experience (choirs) there’s not so much difference below p whereas mp-mf-f-ff is still quite distinguished (fff ends up being just shouting), and you can easily add vocal accents by means different from mere volume. This cannot really be mapped to MIDI velocities with the current dynamics levels / system.

In reply to by Arpicembalo

Sounds like some additional control of articulations is coming for MuseScore 4, see the various discussions in the Announcements forum. And yes, the issue where instruments like piano that are incapable of producing a loud attack followed by an immediate decay - the way it works for almost all other instruments - will hopefully be special-cased, so accents on piano will sound as good as they already do on all other instruments.

Regarding marcato, I understand you are annoyed about the decision to use the common jazz interpretation (short and somewhat accented) instead of less common classical one (same length as sforzato but louder), as presumably your interest is more in classical so the jazz interpretation might not be useful to you personally. But it doesn't help your cause to not at least try to understand why the decision was made, nor does it help to be insulting to those of us who wh dedicate our lives to providing this free resource and to making it the best we can and serve the needs of the most musicians we can. Please understand these really are sound decisions not taken lightly. If you wish to make requests, please do so with understanding of the technical issues as we have carefully explained, refrain from personal attacks, and then rational discussion can continue.

In reply to by Arpicembalo

Calling something an "objective flaw" and "nonsense" and saying you are "tired of people trying to excuse these poor decisions" and demanding we "Fix these annoyances" is inappropriate in any rational discussion of technical tradeoffs. As explained, this was a perfectly reasonable choice given the circumstances. Feel free to acknowledge that you understand the tradeoffs (which your previous statements did not) and why the decision was made, and then offer logical counterarguments that respond directly to the original rationale, if you wish to change someone's mind about the decision or to offer alternatives. But calling their decision-making "poor", their decisions "nonsense". and their explanations "excuses" is a non-starter.

Lots of things may become possible with MuseScore 4, and having rational discussions about the possibilities will be welcome.

In reply to by Marc Sabatella

Sorry not sorry, it's simply the way it is. I've never met anyone so adamant on defending unorthodox interpretations in a widely-used notation program, to the point that you even see your opinion as "the more common one" when that is just not the case. Until you can cite anything but personal experience I'm viewing these decisions as entirely arbitrary.

In reply to by Arpicembalo

I didn’t bother posting any of the zillions of sources that would all universally support what I am saying because you wrote “don't cite jazz rules to me, I know them”.

Again, if you’re willing to engage in a rational technical discussion rather than being rude and insulting, let us know.

In reply to by Marc Sabatella

You didn't bother posting these sources, or don't you need a reference "to prove what decades of experience tells you is true"? "Ask any jazz musician"?

Go ahead, show me the zillions of sources that tell you the "marcato=short" interpretation is MORE common than the classical way. I await them eagerly. Though most like I'm going to get another "Be rational" instead.

In reply to by Arpicembalo

Do a Google search on “marcato jazz”. You will see every single hit explains the “short” interpretation.

As for the significance of how common it is, please reread my posts. What I said was - and I quote - “ MuseScore's playback is based on the genre in which the symbol is used more commonly. It's a relatively rare marking in classical marking, but pervasive in jazz. So the jazz interpretation won out.”

To validate this statement, choose any 10 random pages from 10 scores in the
Norton Anthology (or your favorite collection of classical scores) and count the number of marcatos. Now do the same for Inside the Score (or your favorite collection of jazz scores).

So as explained previously, it’s a judgement call - whichever interpretation we chose, it’s convenient for one group, inconvenient for another. Since marcato comes up much more often in jazz, that was the deciding factor.

In reply to by Arpicembalo

@Arpicembalo:
You can set your preferred values in the articulations section of the mscx file and this is retained when you resave as mscz. This means that you can set the accent levels on a per score basis. I have reduced it from 150 to 120 in a recent guitar score.

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