The direction of a note's stem is based on the number of voices in its context. If there is one voice, notes written above the middle staff line (B in G-Clef) are down-stem, whereas notes written below the middle staff line are up-stem. If there are two voices, the upper voice takes up-stem notes, and the lower voice takes down-stem notes. It is therefore important to determine how many voices there are.

Consider a phrase with two voices, spreading over multiple bars. In one of the bars, the upper voice is silent, so in fact, the bar consists only of the lower voice. Do we consider this bar to contain only one voice, since the upper voice is silent, and then the notes should be noted based on the one voice rules, or we consider it to contain two voices, and the notes should be down-stem (with a full bar rest for the upper voice)? In other words, is the number of voices determined relative to the bar or to the phrase?

To illustrate, option #1:

Example 1 - two voices

Option #2:

enter image description here

  • 1
    Your second picture is wrong no matter which standard gets used - there is no upper voice in Bar 48 there, and the C octave is down-stem when it should be up-stem like the others.
    – Dekkadeci
    Aug 26, 2021 at 12:03
  • @Dekkadeci - Interesting. This is the default behavior I get from MuseScore. Maybe the rules for octave stems are different? Aug 26, 2021 at 12:10
  • 2
    Note that none of the default behaviors in any scoring software should be interpreted as correct notation. They might be correct 95% of the time, or they might be the best way to typeset things on the first pass, or they might be that way because of difficulties in one or more algorithms. The point being, you can’t trust the software to get it right for you. Aug 26, 2021 at 16:34

1 Answer 1


The standard I'm used to seeing is that, if there is a rest in one temporary voice (i.e. you're not writing a fugue or anything polyphonic, so you have no obligation to provide whole-bar rests for voices), you omit the rest for that measure and notate that measure as if that voice is missing. It doesn't matter if that rest occurs in the middle of a phrase.

  • So now I have two conflicting answers and I don't know what to do :( Aug 26, 2021 at 12:18
  • 1
    This is the correct answer. In looking through some scores to find examples for my answer, they seem to follow the pattern you describe: A Bach fugue (e.g., WTC II, Eb major) maintains the stemming (and rests) even when a voice drops out for several measures; but an arrangement of Gershwin's piano solo arrangement "An American in Paris" only maintains the stemming during the measure that explicitly contains multiple voices.
    – Aaron
    Aug 26, 2021 at 12:21
  • @Sipo I've deleted my answer. This is the correct one.
    – Aaron
    Aug 26, 2021 at 12:21
  • @Aaron - Thanks. Just to be sure I understand, you found examples of both patterns? Aug 26, 2021 at 12:25
  • 1
    @Sipo Because my post only accommodated one way of handling the situation, and to expand it would have duplicated what Dekkadeci wrote here without really adding anything. Examples could just as easily be added here. Though if Dekkadeci isn't so inclined (are you?), I could rework my original one as a supplement to this one.
    – Aaron
    Aug 26, 2021 at 17:55

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.