Extreme slowdown on Musescore 2 (and "nightly" 3 as well)...

• Oct 28, 2017 - 02:49

Hi everybody, and thanks to Musescore team guys for their work.
Nonetheless, I'm about to submit another unpleasant thread...
I'v already written about a major "slowdown" five months ago (https://musescore.org/en/node/204806#comment-721546); now I find myself in the same position after months of nearly "quiet" work with Musescore 2. With small scores (e.g. SATB arrangements etc.) I've noticed only minor issues (especially with "complex" lyrics lines and some particular editing), but I'm facing an extreme meltdown with the score uploaded right below... It's a "medium" concert band score (barely 30 pages), through wich almost every kind of operation seems to be extremely slow (safe - ironically - for the navigation/scrolling-down). I'm going really mad: I cannot switch back to Sibelius overnight (I don't want to) but I cannot work out this situation as well ("parts editing" would imply hours for each instrument): can You help me? I'd really appreciate...

N.B. I tried the same score with the latest "nightly" version, without any significant relief. (I'v also deleted from the file a complex piano cadenza that could be a possible slowing down cause...)
P.S. Sorry 4 my English.

Attachment Size
Rapsodia in blu ALTRA PROVA con parti.mscz 522.14 KB


Your score is being slowed down by an abundance of slurs. For example, the Oboe has 5 slurs on each slur in measures 68-70. The slurs on the alto and tenor sax as well as the horn have multiple slurs in these measures. I haven't looked at every measure. I only looked to confirm this is the problem. I have heard of a way to delete the excess slurs, but I don't know what it is. I have never done it.

The first clue I had that the slurs might be a problem was the size of you score. Since it's not a 400 measure 40 instrument symphony it should be half that size at the most. Once someone removes the slurs, you will see how small it should be. I would suggest you wait for someone to remove the slurs for you, because even if it doesn't get fixed for 12 more hours, it will be a lot easier than manually removing every extra slur in the score.

A multitude of notes that are entered do not normally slow down MuseScore like you are currently experiencing. So the piano cadenza, as you figured out already, was not the problem.

I wish there were a way to figure out what causes this. I have never experienced the multiple slurs and I don't know of any of the regulars in the forums that have had it happen to them. Until we figure out why this happens to some people it cannot be fixed.

P.S. your English is excellent.

In reply to by mike320

Thank You, mike320, but I don't see any kind of improvement, even if delete - just to give it a try - every slur on the score... There must be another way to explain this abnormal behavior... (You're perfectly right, here: "Since it's not a 400 measure 40 instrument symphony it should be half that size at the most.")

N.B. For this reason, I'd like to stress out the fact that my pc setting was never an issue - Intel Core i7-4710HQ CPU @ 2.50GHz, RAM 8,00 GB (usable 7,98 Gb) - and, for example, the much bigger recording app "Reason" runs with it without any sign of latency or slowdown...

In reply to by Guido Alici

I tested the theory in the older thread you mentioned an changed all the fonts to FreeSerif. This seemed to have little affect on the speed. I didn't realize the parts were all extracted, so I also deleted all of the parts, which is known to slow down the score. This greatly decreased the size of the score, which I knew but did not think about before. It sped it up a little on my system. I have much less memory than you do, so I'm not surprised I only noticed a little speed up. I am also in great need of restarting my computer, since I don't normally turn it off. I suspect I will see a speed up on this score when I do that. There is still something in the score slowing it down. I have a 208 measure 43 staff score open at the same time and it is much faster than your score. I will continue to look for something odd to see what might be slowing it down.

I just looked at the score in the other post and slurs are not a problem there. I'm not sure why it slowed down. It's not big enough for the size to slow it down like that.

In reply to by mike320

First, the attached file (so, this one: Rapsodia in blu ALTRA PROVA con parti.mscz ) crashes quite easily.
1) Press"I"
2) Remove the two first instruments ie Flute and Oboe-> Ok
3) Undo -> Redo
Result: Crash

I found the cause (issue soon reported)

Not sure right now if this issue/ crash is related to the slowness, but this may not be impossible, because this issue involves several important factors (hide empty staves, and small staves) having a fine, but direct, impact on the layout.
And so, if we add the parts, in a score of 30 pages, it would involve many more calculations, which would explain this slowness. At least, this is my understanding of the problem for now.

In reply to by mike320

Continuing to look at this topic, I observe that if unchecking "Hide empty staves" and having a usual size staff for the last instrument (pianoforte), it does not change much about the fluidity of navigation.

On the other hand, the progress is clear by changing in Style -> Text, the text font of all items, ie FreeSerif instead of MS Shell Dlg2.
You can see this improvement for example by entering notes in the main score, top staff (flute), in first measures.

So, with MS Shell Dlg 2 (initial score): Rapsodia in blu ALTRA PROVA con parti.mscz

And with FreeSerif: Rapsodia in Blue Free Sérif.mscz

In reply to by cadiz1

I see: I can sense the improvement... Let me ask You one thing: is it possible to change all the text style at once (from FreeSerit to - actually - mine - uploaded - Adobe Garamond Pro), besides the manual way (i.e. changing every entrance, like Title, Subtitle etc.) and without uploading a new loaded style???

N.B. After all, I don't want to abandon entirely the possibility to use this Garamond font, which I prefer...

In reply to by Guido Alici

Wait a minute: I see just now that You have deleted all the piano part!!! That's way the speed encreasing... I've just said lately: unfortunately it's not a matter of text fonts... Mike320 wrote it too: "I tested the theory in the older thread you mentioned an changed all the fonts to FreeSerif. This seemed to have little affect on the speed."

In reply to by Guido Alici

"Wait a minute: I see just now that You have deleted all the piano part!!! That's way the speed encreasing..."

Yes! It was my next reply. In my tests, I had observed a difficulty with this piano part (which took me on the way to a bug, now reported).
And I attached by mistake in the previous comment one of this test files (so, without content in the piano part)
So, indeed, the improvement is probably not due to the text font, but more with the content of this piano part (?)
The file with the content of the piano part: 2 Rapsodia in blue FreeSerif second test.mscz
Need further investigation I fear! :)
But we are progressing...

In reply to by cadiz1

I did a test on the piano part. I searched the XML for measures that had the command < tick > (minus the spaces) in them from measure 19, where the piano starts playing to measure 122 only in the piano part. I removed all notes in right hand in these measures and noticed a significantly better response on my computer. I didn't check response after each deletion to narrow it down, because I feared I wouldn't notice a difference from one deletion to the next, but after about 30-40 deletes the difference was noticeable. In the XML, < tick > indicates either there are 2 voices, or a tuplet that has been rounded. I believe only a couple of the measures I cleared had these tuplets in them. Some of them did have cross staff notation as well as 2 voices. These are possible avenues of investigation.

In reply to by cadiz1

Thank You cadiz1, and thank You mike320; I'm really sorry: frustration is speaking out for me... Your last file seems to work slightly better than the original of mine, despite the piano part. Nonetheless, more and more I am convinced that the real cause and reason (for the slowness) just dwells in the piano part (even though I had momentarily deleted some complex "cadenza" bars from it)...

In reply to by Guido Alici

"Nonetheless, more and more I am convinced that the real cause and reason (for the slowness) just dwells in the piano part (even though I had momentarily deleted some complex "cadenza" bars from it)..."

There is something weird about this cadenza in piano part. You can observe (in main score eg) that all the notes are not highlighted when you select the measures 176 to 179. You can delete the first content, but not all the notes.
measure .jpg
To achieve, you have to make the selection from the measure 174 (where appears the text "Cadenza"), or make the selection of two staves (treble/bass clef).

Seems a bit better after the deletion of this entire sequence of measures:
3 Rapsodia in blue FreeSerif cadenza.mscz

But this file has probably suffered.
I don't know what you did exactly for entering this cadenza, you wrote yourself "complex "cadenza", but it would probably useful to detail this.

In reply to by cadiz1

Dear cadiz1, I've uploaded the score with the piano cadenzas some messages ago, do You remember? I re-upload here, one more time for You... As for "lines", I used an extendable one to compensate the tuplets limits, linking that way two 16 notes without beam: look closely, please...

Attachment Size
Rapsodia in blu.mscz 344.65 KB

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