Amplify the volume when two voices have a same note (for non-keyboard instruments)

• Jul 15, 2019 - 11:06

I understand that when two voices have a same note on the piano the volume shouldn't be amplified, but when this happens on wind/bowed string staves, it usually means there are two instruments (or two strings) playing the same note, and in this case the volume should be amplified.
This case seriously affects the playback of the cases in which two brass instruments of the same kind, horn for example, are playing fortissimo while woodwind and strings are also playing in a big volume. Even if others are playing loudly, two unified horns can undoubtedly overwhelm them in a real orchestra, however in MuseScore playback the horns are barely heard because they have the same volume with a single horn.
I know this is only playback-affected and improvement in playback won't occupy the principal part of a notation-software's development, but I still hope that you can take note of this...


This is esp. annoying when those 2 voices (or even different staves, but same instrument) are unison only on some notes, there's a bad drop in the precieved velocity

In reply to by Howard_C.

ISTR (need to verify again) that if 2 staves use same instrument (same sound in Mixer), it gives the same drop in velocity on unisons as it does for voices of a single staff.

Listen to the attached, indeed at least worse with a single staff

Diffcult to implement differently though, as it depends on the instrument, or better, whether a staff contains notes for different instrumnts or not, which nothing MuseScore can really know.
We'd need a way to differentiate instruments by whether they are capable of playing multip pitch at the same time or not, and still would likely get it wrong in some corner cases (think violins, usually just one note per instrument, but can play 2 or even more).
So a clean solution might be to make it s staff/part property with a sane default depending on instrument.

Attachment Size
Untitled.mscz 4.51 KB

In reply to by Jojo-Schmitz

  1. It's indeed better to add a custom option, but default should still be ×n for non-keyboard and ×1 for keyboard instruments. Your note about strings is reasonable, but the notation style for multiple voices on keyboard and non-keyboard is usually different, so even if we want to play the strings in a non-×n method we don't notate in that way...
  2. I don't have computer now so I cannot check the score...

In reply to by Howard_C.

Well, if a staff for, say, flute contains more than note at a time (either as a chord or as separate voices), this got to be 2 defferent instruments, right? So some instruments are monophonic by nature, ather are polyphonic, and that might be recided in instruments.xml for example and be used to feed a staff setting.
And it is not just keyboards that are polyphonic, also guitars, harp (all plucked strings actually), glockenspiel, marimba (all percusive instruments basically) and many more. Human voice, Woodwinds and Brass are all (?) monophonic, bowed strings are in between (leaning towards monophonic).

See above how to get a decent default into a staff propertiy setting. That is needed, because the distiction between monphinic and polyphonic is not always as clear as it seems at first, liek the mentioned exampole of a violin, which usually is a monophonic instrument, but with certain advanced (?) playing techniques can produce 2 or even 3 and 4 notes at a time (the easiest case would be pizzicata on several strings at the same time, relativly easy to bow 2 strings at a time).
I don't see wherere the style for notating multiple voices would be different depending on instrument. Except in the case where the rhythm is identical and voices are not strictly needed but chords are used instead. Doing that however pretty clearly indicates that it is just one single instrument, so nothing the playback would need to consider raising velocity on unisons?

In reply to by Jojo-Schmitz

Firstly, I don't see (actually hear) anything wrong with same notes in different staves, I'm using a soundfont which is more sensitive to volume changes, and measure 2 sounds louder than measure 3.

OK, now I'll explain it.

Let's just take one example:
批注 2019-07-15 235114.png

This kind of notation is already weird enough to appear in any non-keyboard staff. But let's just assume that it takes place on violin's staff. How will you interpret it?

There're two ways:
1. To think of the D in voice 1 as played on D open string, and voice 2 played on G string. This is the common way of understanding it, in this case on the first beat there're two strings playing D, so the sound of D should be amplified.
2. To think of only D is played at the beginning, with a single string, and later A is added by playing on G string. Do you really need to notate it this way if it means this?

So this is the point -- to demonstrate a contrapuntal sentence/paragraph clearly on score, this kind of notation is needed, and only in this case should the playback not be amplified, because the two voices are just written representation of two co-existing melodies, and don't stand for two instruments, or two same parts (like two strings).

But for non-keyboard instruments, there's a very few ways that notation can be like that, and even though it's like that, it shouldn't be interpreted like keyboard instruments. For other instruments (except harp), it's impossible that a counterpoint can form within a single instrument, so the explanation is either multi-instrument or multi-string. (Regarding harp, there's something special, if you are a harp player, or have studied orchestration, it's common sense that a harp can play the same black-key note on two adjacent strings if you set the pedals in a certain way. In this case the playback should be amplified, but again, the software doesn't do it, for example if you input D-sharp and E-flat together.)

The basic difference is that you can hold keys while playing other keys, and still make the holded keys belong to different voices on a keyboard instrument, which you cannot do on all other kinds of instruments (even if you can on strings, the effect that pressed strings can sound like part of two voices definitely isn't obvious, so we don't write string staves like this. And for harp, I don't read many harp staves so I don't know, but I feel that this way of notation isn't popular when it comes to harp).

In reply to by Howard_C.

If I saw this on an instrument that normally plays single notes (including strings) I would look for the divisi or a similar indication telling only some of the musicians to play the stem up notes. If that is missing I would try to figure out where it belongs. If this is the only note with two stems I would consider it a typo. I would only interpret it as one musician playing on two strings if it is clearly indicated to be the case.

In reply to by mike320

Well of course it can be explained as divisi, in this case the volume should be manually discreased. This is a special case for strings, I think if this issue is taken care of in the future, we can add a text property for decreasing the volume in case that text is “divisi”. Then ×2÷2=1! The volume of D on first beat remains the same, as all members of the strings are playing it. (Just using 2 to simplify the case here, it's not 2 anyway)

In reply to by Jojo-Schmitz

I don’t think the underlying assumption in the OP’s post is correct. While two, for instance trumpets, are louder than one trumpet, they are not twice as loud. Even if playing at exactly the same frequency, their respective waveforms are not synchronized, which would be required for a doubling of amplitude.
This is not to say there isn’t room for improvement in this area, obviously there is.
The issue becomes more complicated when considering other instruments. While two trumpet players will have a harmonic profile that are very similar, it isn’t necessarily the case with other instruments. The human voice is probably the most extreme example, with well-trained voices having at least as much acoustic energy in the harmonics/formants as in the fundamental, but still a different harmonic profile from each other and from less well trained voices.
Again, I would also like to see improvement but there is no one-size-fits-all instruments fix.

In reply to by marty strasinger

This is very true - it's not twice as loud in any literal sense but something more logarthmic. Also, it wouldn't be a change in velocity per se but in volume for the same velocity. That is, there is a significant difference between the sound of one trumpet playing loud and a bunch of trumpets playing quietly. You can add enough trumpets to make the volume the same, but the tone quality is still that of trumpets being played quietly, and this is lost if you use velocity. Luckily, with single note dynamics, that's not so much of an issue - we no longer rely on velocity (alone) to determine volume.

The real problem currently are the phase artifacts that happen currently because of the same exact waveform being played twice simultaneously. I wonder if merely adding some randomization to the CC messages for volume would fix that?

This is a hideously complex problem which also affects closed-score choral scores (e.g., hymns). MIDI channels and MuseScore expose a "keyboard model", where any single note can either be on or off. For multi-stringed instruments where multi-string unisons are possible, this just doesn't work, or for putting multiple "desks" on the same channel. I think the world understands and expects that MIDI scoring works this way. If there is ever a feature that attempts to compensate unisons with volume and/or detuning, this had better be highly optional per staff.

In reply to by BSG

  1. +1 to BSG’s last post.
  2. To mike320’s idea about using two different soundfonts: brilliant! Wish I had thought of it (with Oscar Wilde saying “you will, marty, you will” in the background).
  3. Clumsy workaround- manually change the dynamic level of the note(s) of concern.
  4. More science/physics stuff: given how technical this discussion is, the term “volume” is not appropriate. “Amplitude” is the correct term.
  5. Over my head/needs someone smarter than me: given the non-linear response curves of the human auditory system (to both frequency and amplitude), and that changes in amplitude are not linear with changes in perceived decibels (as Marc said, it’s logarithmic) I can only assume this would be a ludicrously complex project. If anyone has a less clumsy workaround than I suggested, would love to hear it.

@Howard_C... You wrote:
I understand that when two voices have a same note on the piano the volume shouldn't be amplified, but when this happens on wind/bowed string staves, it usually means there are two instruments... playing the same note, and in this case the volume should be amplified.

Here are some examples showing how slightly de-tuning can be effective. (Created/tested with MuseScore's default soundfont):

Two trumpets, separate staves, playing in unison:

Multiple trumpets, separate staves, playing in unison:

Two trumpets, same staff using voices, some unisons present:

My examples are from this past forum discussion:


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