Friday, September 21, 2012

MuseScore vs. Score vs. LilyPond vs. Sibelius

Some years ago I came across a really interesting essay that compared Sibelius' output to the good old, and by many music publishers strongly preferred Score application. You can read it on Jeffrey Grossman's site.
At that time I was a devoted LilyPond user, so I replicated the score with LilyPond to see the difference. And now as I'm getting deeper and deeper into MuseScore I did the same with MuseScore.
So I have the same score engraved with all four applications - see the differences yourselves.



The Score and Sibelius output is directly from Jeffrey Grossman's site (thank you a lot!), so I don't know if there are any tweaks, nondefault spacings etc. set up. LilyPond engraving was done with LilyPond 2.12, MuseScore is version 1.2.

The Score output is quite good, I would say it is superior.  The Sibelius problems are described on the original page.

The LilyPond output is good as well, the spacings are ok, though there are some minor problems:

  • the most severe is 16th beam-accidental collision in measure 3
  • the hairpin alignment (can be easily fixed though)
  • the > hairpin start position in measure 3
  • the ties in measure 3 (though I think that would need tweaking in all softwares)
 And finally, the MuseScore format seems to have the same problems as the Sibelius, so some note positioning tweaks are needed (besides all the stuff you have to move manually at the first place, dynamics, hairpins etc.)

Score 

Sibelius

LilyPond

MuseScore

Finally, here you can see the somewhat tweaked MuseScore output and download for your own experiments from musescore.com

MuseScore "Triplets" test by briff

9 comments:

  1. New LilyPond versions don't have the collision between beam and accidental.

    Agreed on other stuff, though - needs fixing.

    ReplyDelete
    Replies
    1. Has the hairpin bug been filed? I cannot seem to find it in: lilypond hairpin bugs

      It would be nice to get the result with latest 2.17.20 LilyPond up, 2.12 that's from 2008?

      What's the problem exactly with the ties, can we get a bug report please?

      I would like to note that LilyPond is a compiler and therefore ask:
      are the other scores (Score, Sibelius, MuseScore) the default untweaked
      output from those softwares? Or are we comparing the LilyPond expert-system here against a human typesetter (who uses a dedicated drawing program?)

      Delete
    2. Two ties are a bit too close.
      However, it's a bit hard to fairly compare as some graphic elements like crescendo pins must be manually drawn so its vertical position is never automatic. But besides that the default untweaked output is shown in MuseScore, but I'm not so sure about Sibelius and Score, as I didn't create those.

      Delete
  2. Hi, I was looking for online course on music programs through which I can improve my knowledge in music and music softwares. I found this online course : http://www.wiziq.com/course/3200-master-sibelius-7-software
    Help me by reviewing this music software course, so that I can make a decision to join it.

    ReplyDelete
  3. Hi, nice example... We wrote an similar comparison on the German Wikipedia article. Instead of LilyPond, we also tested PriMus, Finale and Amadeus... You can have a look on http://de.wikipedia.org/wiki/Notensatzprogramm#Programme_und_Anwendungsbereiche

    ReplyDelete
  4. What's with the note stem for the rest in the right-hand triplet in bar 3? IMO, only Sibelius gets that right.

    ReplyDelete
  5. I'de like to add the example in my favorite notation program. Most off positions (including slurs, ties, accidentals, stem lengths, beam angles, ...) are default settings without manual editing.

    (exeptions: manual editing in measure 3 for the position of the triplet line and in measure 4 for the tremolo position)

    pdf and png can be downloaded here:
    https://dl.dropboxusercontent.com/u/5282959/PriMus%20Triplet%20Test.zip

    ReplyDelete
  6. Musescore's error about the triplets in measure three has been fixed.

    ReplyDelete
    Replies
    1. Which error? The first triplet is not an error, that's manually tweaked, though it would be good if MuseScore supported that natively.

      Delete