Slur Lines - Buggy or Sloppy
Hi there,
So a while ago one of the admins asked me to make note of slurs and how they were working as sometimes they don't quite go in the right spot when they are assigned. I've been saving some examples so they can be highlighted so as to suggest or inspire possible corrections. Some of the examples might not be perfect as the new version made some of the formatting...different, but anyways...I've separated the matter into two categories, and I'm able to come up with two culprits.
1 - Buggy Category: This is when the slur clearly isn't connecting the starting note and the end note properly. Usually the result is that the slur is floating oddly above where it should be.
-The main culprit for Buggy Category would generally seem to be connected Eighth notes somewhere in the slur.
2 - Sloppy Category: This is when the slur either crosses through a note's stem or crosses past the "dashes" in lyric lines, OR if it simply come too close to either of these (I think anything less than a half centimetre of space between slur and stem or slur and dashes looks really, really bad).
-The culprits that cause these are either a LONG string of notes that are all slurred together, or a short set of notes slurred together with a particularly low (in Bass clef) or particularly high (in Treble clef) note as the centre of the slurred notes.
I've included the scores for my examples so far as attached files. Below are descriptions of where the aggravating slur can be found in the score (I always find a way to correct these manually).
Example 1A - See measure 4 - the slur clips right over the lyric dashes.
Example 1B - See measure 4 - the slur clips too close to the lyric dashes (the distance between the slur and the THIRD dash in that measure is roughly what I expect to be the minimum distance for it to look presentable).
Example 1C - I think I saved this one as an example of one that doesn't look sloppy in measure 5, although I think even a fifth of a centimetre more would make it look even better yet.
Example 2A - Two examples here. Measure 2 has a bit of a "floater" in the bass clef at "mercy," and measure 4 again has the slur crossing the lyric dashes.
Example 2B - Measure 4 in the treble clef touches the lyric dashes.
Example 3A - Measure 5 has a slur touching the lyric dashes in both clefs.
Example 3B - Measure 4 bass clef slur touching the lyric dashes.
Example 4A - Measure 6 slur is kind of a floater and just doesn't look right. Measure 4 slur comes too close in my opinion to the first note in the eighth-note trio, and also generally looks pretty awkward overall.
Example 4B - Slur in measure 2 treble clef for "el" (in Israel), third note in alto - the slur comes too close to the note stem.
Example 5A - Measure 5 in bass clef - the slur is a major floater.
Example 5B - Measure 3 slur in bass clef for the bass part - the slur comes too close to the 2nd note's stem.
Example 7B - (I couldn't figure out why I saved 6B...) Measure 5 - slurs are coming too close to the lyric dashes.
Most of these are of the "sloppy" kind, although I will definitely encounter more of the "buggy" style as I get more into the eighth-note musics as I continue my projects, which will cause a serious amount of certain hits.
As a bonus buggy feature, I've included two more scores, "Reverse Run 1" and "Reverse Run 2." In measure 3 of both of these scores on the last note of that measure (the half note), the playback highlighter curiously plays backwards on the last note, even though no visible effect occurs in the audio. I'm baffled by these and these haven't been the only ones. What I can say is that they seem to show up after using the "split measure" option on a note in the proceeding measure (that is, measure 4). I often do this to add or delete a single note, and then re-join the measures, and this causes the problem.
Cheers and let me know what other particulars you'd like me to look for, or if you want instead to have examples of how I modify my slurs to be the nice way.
Good day!
Attachment | Size |
---|---|
Example 3B.mscz | 19.66 KB |
Example 4A.mscz | 19.91 KB |
Example 4B.mscz | 8.29 KB |
Example 5A.mscz | 16.68 KB |
Example 5B.mscz | 25.45 KB |
Example 7B.mscz | 25.28 KB |
Reverse Run 1.mscz | 22.99 KB |
Reverse Run 2.mscz | 22.09 KB |
Example 1A.mscz | 26.91 KB |
Example 1B.mscz | 19.47 KB |
Example 1C.mscz | 24.76 KB |
Example 2A.mscz | 24.4 KB |
Example 2B.mscz | 20.07 KB |
Example 3A.mscz | 23.22 KB |
Comments
Thanks for going to the effort of collecting and posting these!
One general thing I can tell you is that lyric lines (dashes, melisma underscores) do not participate in automatic placement at all. That's a known limitation, #280139: Autoplace fails to avoid melisma underscores; it isn't specific to slurs. That explains a number of your issues.
For the rest, I would say these things are subjective, and while we do our best to produce results we think will be generally pleasing, there is going to be no getting around the need to adjust some things manually to get the results you personally might subjectively prefer. There should be far less need in 3.1 than in 2.3.2. I just set a big band arrangement (120 measures, 17 parts) and needed to adjust only a four or five slurs, which would have been unheard of previously.
For the record, the way the algorithm works is, we first set the slur based on the default settings, and if there are no actual collisions, we do no further adjustment. So that's the cases where you may see things being to close. Due to the data structure involved, it would be a lot slower to try to enforce a certain distance if there no collisions. If there is a collision, then we move it to avoid the collision, and add padding as per Format / Style / Slurs / Autopace min distance. So the ones you call "floaters", these could be brought closer as a rule by reducing that parameter. Still, I think what you'll find is that some of these appear to be floating higher than the algorithm things they are, because most elements are considered to have a "box" around them, and we are avoiding the entire box. Like the one in bar 5 of example 5A, draw a box around the flags on the sixteenths and you'll see where we are actually measuring against. Doing better would require the font designers to provide more detailed metrics on their glyphs. We actually do this for accidentals, so they can stack tightly on chords, but that's about it.