Bug? Some measure Repeats ignored first time through bar repeat?
Hi Folks,
The first time through the four bar repeat (13 to 16), the last two measure repeats (bar 15 and 16) of the Cabasa and Triangle are ignored. I tried restarting the program, the audio driver and reinserting the measure repeats--still no luck.
My system: OS: Windows 10 (10.0), Arch.: x86_64, MuseScore version (64-bit): 3.1.0.7011, revision: 8b1a7dc
| Attachment | Size |
|---|---|
| Xote_Swing_arrangement.mscz | 22.16 KB |
Comments
Also Surdo 1 in that part, the second groove repeat (13 to 16), is much more quiet than in then in the first Groove repeat (Bar 5 to 9) part? I've checked the inspector and the velocity is set to default? It has something to do with previous forte? But, the other instruments are not effected, at least to my ears?
I have looked at this. I reentered the measures that won't play right and they still don't play. I also looked around for some misplaced item that could affect the playback and didn't find anything obvious. I did confirm measure 14 plays fine and it's identical to the two that don't.
In order to understand the problem, I tried to export it as musicxml and then import it again.
But all the bar-repeat signs are gone. : /
In reply to In order to understand the… by Ziya Mete Demircan
That sounds like a musicxml problem in need of fixing. I unrolled the repeats and musescore seems to do it properly, so Musescore sees them.
Is there something in those measures that might otherwise affect playback?
In reply to That sounds like a musicxml… by mike320
found 2 strange staff text (x offset, with 86.xxx value) but those no affect to playback.
in mscx: "master" , "fluid", and "zita" definitions is x 2 (not duplicate; separate)
In reply to found 2 strange staff text … by Ziya Mete Demircan
I think:
all the scores transferred from 2.x to 3.x has some tags defined twice in the "synthesizer" and "articulation" definition sections.
What I found in the articulations: "normal" (no name=""), tenuto, marcato, sforzato
In reply to I think: all the scores… by Ziya Mete Demircan
Thanks for looking into this. This score was completely created in Musescore 3.1. I did use a custom instrument.xml (see attachment) for the instruments in the score, in which I redefine some of the articulations. However, replacing the non-playing instruments with the same instrument.xml, but without the articulations redfined, did not fix the issue of repeats not playing first time through. On a side note, removing the dynamics (piano and forte) in the previous measure to the 4-bar repeat does help restore the volume/velocity of Surdo 2 (second unrelated issue) but had no effect on the repeats. Also, reentering the notes of the Cabasa, before the measure repeats, did not fix the issue. Oddly, the first four-bar repeat of the Groove in the score is an exact duplicate and it works fine. I double check this copying and pasting the second four-bar Groove repeat into the first 4-bar Groove repeat, replacing it completely. There is something about the positioning of the second Groove 4-bar repeat that is different.
Brazilian_percussion_with_articulations.xml
In reply to Thanks for looking into this… by Sambaji
The measure repeats not playing first time in the 4-bar Groove repeat has something to do with the bar 15 position of the 4-bar Groove that is the issue. Removing measures 9 to 13 of Break 1 does resolve the issue. Reinserting 4 new measures (9 to 13), restoring the 4-bar Groove to the bar 15 position, returns the issue.
In reply to The measure repeats not play… by Sambaji
When you say restoring the 4-bar Groove to the bar 15 position, are you deleting the measures with ctrl+delete then inserting new measures using insert (or ctrl+insert) then reentering the notes by hand, or are you copy delete insert and paste the notes?
If this were imported from an earlier version I would blame it on that. This is made from scratch in 3.1 and is a problem that needs to be understood and fixed.
In reply to When you say restoring the 4… by mike320
I meant restoring the Groove to the Bar 13 position by deleting and reinserting measures 9 to 12 with ctrl+delete and then "Insert Measures" under the Add menu. Although the score was created by scratch, I copied some of the notes from another score. To simplify the diagnosis, I deleted as many elements possible where still retaining the issue--for example, removing all but the cabasa instrument and the first Groove repeat. And yes, the problem persists. See single instrument version of the score. I was mistaken about Bar 13 being relevant. At Bar 9, the problem persists. Keep in mind that the triangle was also having the same issue so it's not instrument specific.
In reply to I meant restoring the Groove… by Sambaji
I've narrowed the issue down even more. If I delete the first measure, the problem goes away. If I delete any item in the first measure, the problem persists.
I don't see anything in the .mscx file that I would be deleting with the first measure that I can't delete individually except the time signature. As a test I created a 4/8 time signature (to keep measures in tact), inserted that and the problem persisted.
In reply to I meant restoring the Groove… by Sambaji
What's particularly odd is that measure repeat in bar 10 is played, but the repeats in bars 11 and 12 are not first time through. Looking at the musicxml (in the simplified score), the core code for bar 10 is essentially the same for bars 11 and 12? The issue persist even after removing the ending barline repeat.
EDIT: As Ziya pointed out the repeats are not included in the musicxml. The repeats are missing in the reimported musicxml version. Yet they are present in the mscx file.
In reply to What's particularly odd is… by Sambaji
There is something, seemingly unrelated, in the score causing the problem. On many occasion @Cadiz1 has found these items. His score diagnostic skills seem magical at times. I've seen little of him in the past week or so, perhaps he can look at the score or possibly a programmer can look at the score and see what variables are changing during playback each time and find the cause of it.
In reply to There is something,… by mike320
I've file the issue in Tracker, attaching the simplified score. #289721: Bug: Measure repeats ignored first time through
In reply to I'll file the issue in… by Sambaji
The failure to export the measure rests is a seperate issue (at first glance) since in measure 10 the measure repeat does not export either, but does play back properly in both versions. The musicxml export is also not a regression since 3.0.5 doesn't export it either. I'll file an issue against that also.
In reply to The failure to export the… by mike320
See #289724: Measure repeats not exported to musicxml for the measure repeats not being exported.
For the playback, Marc Sabatella has a theory and is working on a solution.
In reply to See #289724: Measure repeats… by mike320
Theory, yes, but a dead end so far. Letting someone who knows that code better take it further.
In reply to Theory, yes, but a dead end… by Marc Sabatella
Between Moe, Larry and Curly here, we couldn't find the cause in the .mscz file so I decided it had to be an obscure problem in the program. Since you can't recreate it from scratch, I would say this supports my theory that the problem will probably not be common. I know I'm not telling you anything you don't know, but it will likely drive people crazy when it's found again until it is fixed. I hope it's fixed before the final 3.1, but I would say at this point it shouldn't prevent its release.
In reply to Between Moe, Larry and Curly… by mike320
Deleted.
Shemp Howard.
In reply to I don't know about that. I… by Unknown Prodigy
There are no endings involved in this issue, only measure repeats that look like a percent sign (%), so you are probably seeing a different issue. Lets keep those discussion there for now.
In reply to Thanks for looking into this… by Sambaji
In this case, I am correcting:
If you are using a special "instrument.xml" file, the definitions in this file will be appended to the end of the articulation definitions in the current "mscx" file, instead of replacing existing definitions.