Unusual dynamic changes, why?

• Sep 26, 2018 - 06:11

Okay, so I am composing this piano sonata. I was inspired by Mozart and got the first theme down. I might want to slow the tempo down just a tad. But I notice something that seems odd and it has to do with dynamics.

So I was wanting an effect in the scale passages of like swelling and diminishing but not really going far outside of a certain dynamic. Here is how I did it:

1) Selected Crescendo hairpin and set it to half the measure. Same with Diminuendo hairpin
2) Repeat until I reach the creshendo to forte
Then I listened to it and noticed it got forte right away at the scale passages so I continued.
3)Add piano and mezzo piano dynamics(1 per measure) and set them invisible
This turned out to not be enough so I added more redundant piano dynamics at the end of each measure in the scale passages.

But I still get something odd. It seems to be doing a Crescendo where the Diminuendo hairpins are in measures 6-8 despite my efforts to get the "swell and diminish" effect? Why would it Crescendo to forte where I specified Diminuendo and then do Subito piano where the invisible piano dynamics are? That is completely wrong and yet it is doing it. I was wondering if you could help me correct this and get the "swell and diminish" effect I want.

Here is the score:
Piano Sonata no. 2.mscz

The dynamics lines look for the next letter dynamics to crescendo and decrescendo to. If the final dynamic for a crescendo is lower than the starting dynamic the volume is immediately changed to the new dynamic. Decrescendos work the opposite. In the inspector there is a Velocity change field that will over ride this. Basically Velocity is the term used in Midi for volume. You can control how much each crescendo or decrescendo changes the volume with this and the next dynamic mark will be ignored for volume change for that hairpin. You can examine dynamic marks in the insector to see what their velocities are so you can make reasonable adjustments. By the way, all velocity changes are between -127 and 127.

But how does that help me if I already have the needed dynamics in there and have the ones where the hairpins are set to be invisible and it is still not giving me the "swell and diminish" effect?

I tried setting the velocity change to 15(the difference in velocity between piano and mezzo piano), didn't make a difference. I tried deleting and then adding again the piano dynamics in those measures where I hear a creschendo where the diminuendo hairpins are to see if anchoring them to the last 16th in the measure would make a difference and I didn't hear a difference. Finally, I thought "Maybe the mezzo piano is where the issue lies."

So at first I tried increasing the velocity of the mezzo piano to be more like that of mezzo forte. That didn't work. Then I deleted and added again the mezzo piano dynamics in the affected measures and voila, I am getting the swell and diminish. I guess tweaking with the position and the anchor point also meant that the effect I wanted went away.

Moving dynamics with the mouse can accidentally change staves as well as notes, so best to avoid that. Also, the dynamics have to be before/after the hairpin for the automatic playback to work - overlapping won't work. When in doubt, use the Inspector, but again, 15 is pretty subtle, no surprise you have trouble hearing it.

In your attached score, the difference in volume between p and mp is minimal. A human would make the p softer and/or mp louder to achieve the volume difference, but the computer has a set volume unless you change it.

I hear the swell effect in those bars just fine. it's subtle, because you are going from only p to mp. Change the mp to f and you'll hear it more clearly :-)

But you do have issues where you have some dynamics attached to staff 1 and others to staff 2. In order for the automatic playback of crescendo and diminuendo to work properly, everything needs to be in the same staff.

In reply to by Marc Sabatella

Yep, that's what I realized and fixed. I also added a piano dynamic to the start of the crescendo to forte because without that piano dynamic, even though there was a piano dynamic in the previous measure, it was like "Okay I will play this crescendo starting at mezzo forte(default velocity)" so I added the piano dynamic to the start of that crescendo and set it invisible so that the crescendo would start correctly.

Sorry, but imho it is a bug that a decrescendo hairpin does not end a previous crescendo hairpin. The request of having a letter for dynamic after each hairpin is artificial and widely disobeyed in existing scores. A real musician would increase the volume during the crescendo and decrease it back to the volume before the first hairpin during decrescendo. This is requested quite often with a pair of hairpins.

There is no requirement to have a dynamic marking, it's just one of the several ways to tell MuseScore how much crescendo or diminuendo you want. You are welcome to add the dynamic and leave it visible, to add it but make it invisible, or to not add it at all and instead simply use the Inspector to tell MuseScore how much of a velocity change you want. Having MuseScore pay attention to the dynamic marking is just a convenient way to get reasonable behavior automatically in the most common case.

In reply to by Marc Sabatella

My experience with Musescore 2.3.2 Revision 4592407 under Ubuntu 18.04.01 LTS is the following.
If you have a piano section for quite a while and then insert a crescendo hairpin and after the end of the hairpin insert a diminuendo hairpin then the second hairpin is ignored and the playback becomes louder all the time until a new dynamic text, e.g. p for piano is found. I would expect, that with the diminuendo hairpin the playback would become softer and that is not the case. Am I clear?

You are clear, although actually, the way you described it isn't quite accurate. If the preceding and following dynamic are both "p", then the crescendo will do nothing; it won't get louder.

Anyhow, my point is, regardless of what the default behavior is, it is just that - a default. But like most things in MuseScore, you aren't limited to the default. If you choose not to add a dynamic marking immediately after the hairpin, that is your right. If you wish MuseScore to do obey your intent during playback despite not having those markings those markings, simply use the Inspector to tell MuseScore the velocity change you want. It works perfectly - give it a try! In your example, click the crescendo, right-click the diminuendo, go to the Inspector, and set their velocity change to 40. Now play the passage it will get louder then softer exactly as you instructed MuseScore to do, despite the lack of a dynamic marking.

As an additional feature, some day it might be nice if this particular pairing always set a default velocity change to save those couple of extra clicks. Feel free to submit a feature request in the issue tracker for this!

In reply to by Marc Sabatella

I checked it once more. There is no dynamic after the second hairpin because I am in the process of entering the piano. The inspector reports, as you expect, change rate 0. Nevertheless, the piano gets louder with the first hairpin. The second one effects no change. See measures 74 and 75 in the piece appended.

Attachment Size

Change rate is only different from 0 if you've changed it manually, it is not a reflection of the change rate you will hear during playback caused by dynamics.

As stated before, by default MuseScore interprets the change of a dynamic line by looking at the preceding and following dynamic in the same staff. So let's start with the crescendo of m74:
* preceding dynamic marking is the p from m70, which has a velocity of 49
* the following dynamic isn't found (hairpin of m75 has not yet been evaluated and no dynamic marking follows it throughout the entire score) → special case MuseScore defaults the dynamic to its default level of velocity 80
Result: the hairpin will playback as a velocity change from 49 to 80.
Moving on to the next hairpin (decrescendo at m75);
* preceding dynamic is the evaluated end result of the hairpin just before (80)
* again, no following dynamic is found, so the default velocity of 80 is used
Result: this hairpin will playback as a velocity change from 80 to 80.

Similarly you'll notice that your hairpins for the Clarinet also probably don't play back what you want. Take for example the crescendo at m5-6:
* preceding dynamic of p found at the start of m5, velocity 49
* following dynamic is then found at m21, the pp, velocity 33
Result: in absolute terms, the change would result in quieting down; as MuseScore knows you've entered a crescendo, it'll not result in quieting down, but is floored at no change.
The crescendo in m10 then has exactly the same evaluation: from 49 (end of previous hairpin) to 33 (that same pp) resulting in a 0 increase of volume.

For most users is it often easiest to add a hidden dynamic to force hairpins broadly respond as they react. As Marc said, if you don't want this, then you'll have to tell MuseScore which change it should apply across that hairpin.
See attached where I've done that (a change of 31 each time) for the here mentioned hairpins.

Attachment Size

Thank you for the clear explanation. I thought setting a change of zero is a means to not playback that hairpin. Now I see that this is a mark for calculating the change from previous to next dynamic, i.e. not overridden. Perhaps I have from now on a better control of dynamics rather than just remove hairpins.

To be clear: a setting of 0 means "apply the automatic algorithm to calculate the velocity change based on the preceding and following dynamic marking". You post suggests one simple improvement to that algorithm would be to simply do nothing if another hairpin is found before the next dynamic. You'd still be forced to enter your own value as I suggested if you want it to playback, but at least the default playback wouldn't
be counterintuitive as it occasionally can be now.