Challenge II

• Feb 25, 2010 - 00:08

The first MuseScore Challenge proved to be quite a success: within a week, several new features were added and some bugs ironed out. It seems to be a great way to discover missing features or bugs in the software, so therefore, a new Challenge!

Craig Sapp from CCARH again typesetted a score in SCORE. It's a short but tricky prelude by Johann Nepomuk Hummel, only 7 measures long. The pdf file is attached as a file to this forum post.

It features:

  • multiple voices
  • lots of expression markings and text
  • an overdose of tempo changes
  • a candenza measure that does not adhere to the 4/4 meter (at least not at first sight)
  • septuplets / octuplets. The latter doesn't make much sense in a 4/4 meter, but the composer apparently put them there...
  • staff crossings
  • editorial marks (the natural sign on top of, instead of before, a B in the last glissando means that it was not in the original manuscript, but the editor thinks the composer made an error and it should be there)
  • ...

Give it a try, and if you come across something in MuseScore that doesn't seem right, submit a bug report! Or a feature request, if something's just absent. Who's the first to come up with an exact copy, typeset in MuseScore?

Attachment Size
HummelPrelude.pdf 223.62 KB


After looking at this score there doesn't seem to be anything that can't be done with 4 or fewer voices per staff. I may be wrong since I'm not expert on those sort of things, but I believe 4 voices is a limitation of the MusicXML structure. Someone will surely correct me if I'm wrong.
Expression markings and text:
There is nothing difficult about that, just type a staff text and move it into place.
Tempo changes:
There's not that many here, and only one in any given measure. What MuseScore DOES fail to handle are mid-measure tempo changes. I use these to achieve effects like Ritardando, accelerando, fermatas, pauses, and con licensia. Although the tempos can be placed, when the file is reloaded, they will not play back properly, and I suspect corruption of the file results in some cases.
These are easy enough to do. Just write them out ignoring bar lines, then count up the note values, copy the notes, Change the actual value of the measure containing the cadenza to something that fits (for example the last measure would be 23/16), then paste the notes back into that measure, cleaning up unneeded rests by making them invisible.
7 and 8-tuplets work in 4/4 time. However the sephtuplet uses too many beams. There's a bug there.
Staff crossings:
Are entirely possible, although there is a bug the I have reported when manually beaming a group of notes that crosses between staffs. Also, I am unaware of a method to move a single note or selectable group of notes in one chord to another staff.
Editorial accidentals:
Musescore allows these, however it follows the normal convention of placing an accidental in parentheses before the note. I haven't seen any other instances of this type of editorial marking in a printed score.
On to what really glares at me as being a challenge for MuseScore:
Pedal indications are part of the music. If these continue to be discarded, playback is wrong as all notes will have the wrong durrations. Although Musescore does allow the pedal marks to be placed they are ignored when the score is played back. I'm sure also that many (most?) pianists would prefer the more descriptive line type pedal notation. The look of this can be achieved to some extent by placing lines with hooks at both ends, but the _^_ type of pedal point is still impossible.
Ties such as those in the left hand in measures 1 and 2 are impossible as they are not strictly (anal-retentive-ly) correct notation, even if they are common. You can get the look by using slurs but the sound would be all wrong even if the pedal were used.

In reply to by [DELETED] 5

Thanks for the correction.
By the way, though, what exactly is it that makes this score so complex. It looks like just regular ol' piano music to me. Complexity would come with things like ossia, editorial rewriting of entire systems. grace notes combined with arpeggios, 2nd movements, fermatas..... etc. The only complexity I really see are the ties I mentioned above.

In reply to by MDMilford

Maybe, SCORE users consider it a complex score because SCORE is doing the layout decision himself. MuseScore does not or only very very minimal. In particular, there is no text/note colision detection, no auto vertical layout etc...
An overdose of texts in a SCORE document means a lot of computation to make it right. In MuseScore it means some drag and drop and human intervention.

In reply to by MDMilford

MusicXML has no limit on the number of voices per instrument or staff. However, most software systems do - Sibelius has a 4-voice limit per staff, Finale has 8 (2 voices x 4 layers).

Score is used just for engraving, not playback, so if MuseScore could make it play back correctly that would be a bonus. You can make this play back correctly in Finale with some tricks, but those tricks add to Finale's complexity in a way that may not be desirable for MuseScore.

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