Why musescore.com upload halts playback? Score has multiple sections, repeats, and D.C.'s to end.
Trying to figure out why when I upload the attached .mscz to musescore.com, the playback stops at top of page 17 on second repeat of 4x volta. I believe all repeats and voltas are valid. I have used modified D.C.'s multiple times in the score, but set them all to only "play to end". It plays fine on linux and windows, and also if I export wav or mp3, the generated audio plays to end. URL: https://musescore.com/ericfontainejazz/mega-man-2
I apologize if i've made a rookie mistake, but any help is welcome. I'm aware there are still bugs with repeats.
| Attachment | Size |
|---|---|
| Mega_Man_2_incomplete.mscz | 44.87 KB |
Comments
I don't understand what the expected roadmap is. After the first DC, which says "then Coda", it looks like you have deleted the instruction that actually tells it go back to the beginning, and have set it to *not* take the coda. I'm guessing also maybe you mean for the DC's to just go back to the start of the movement, but that isn't implemented - you'd need to add markers (eg, a Segno, possibly made invisible if you think it will be understood to the human musicians playing the score that "DC" means start of movement).
Also, what version are you using?> There have been changes from 2.0 to 2.0.1 and then in mightly builds since then. This could have something to do with differences you are seeing.
In reply to I don't understand what the by Marc Sabatella
Errors still with linux & win stable 2.0.1 and nightly.
After much investigation, I've discovered the culprit for playback to halt. I was using jump instructions that contained empty strings, which results in playback halting at end of second time though next repeat. I've generated issue report: https://musescore.org/en/node/66946 "Null/Empty/Invalid Jumps should either not interfere with playback or should report error"
I've also bumped https://musescore.org/en/node/16397#comment-305416 "Da Capo ignores section breaks". I will use your suggested workaround to create invisible markers. THANKS!
(No need for me to explain to you my score's roadmap now that I have discovered the problems and know how to workaround.)
In reply to Errors still with linux & win by ericfontainejazz
Could this be #64241: D.S. from within repeat bars to a Segno outside crashes on playback?
In reply to Could this be #64241: D.S. by Jojo-Schmitz
I'm running your jump-crash.mscz in my nightly windows, but in my opinion I don't think it is the same bug.
In reply to I'm running your by ericfontainejazz
Description might be different, but it seems quite likely the same fix applied to both cases. There were other fixes too. In any event, your file works in current builds.
In reply to I don't understand what the by Marc Sabatella
Marc, what is the expected behavior of jumps to markers at greater measure numbers? Is that undefined behavior, or should it playback normally after taking the jump? In example below, after skipping to sign, playback will take the 1st ending of next repeat, but will then skip over 2nd ending without even doing a repeat:
https://musescore.com/ericfontainejazz/behavior-of-repeats-on-forward-j…
In reply to Marc, what is the expected by ericfontainejazz
The standard interpretation of music notation is that repeats are not taken on a DS/DC unless explicitly called for. However, repeats *should* be taken after the jump to coda. MuseScore implements the standard interpretation, but does not currently provide overrides to alter this behavior. See for example #8604: Better control of repeats and jumps for playback.
In reply to The standard interpretation by Marc Sabatella
According to https://musescore.org/en/node/8604#comment-90692, I have added some tests to mtest/libmscore/repeat/ directory for slightly different test cases of forward jumps:
I updated tst_repeat.cpp with these cases, but again please let me know if I've done wrong. I've made pull request for my checkout "66851-forwardjumptest" https://github.com/musescore/MuseScore/pull/2103
In reply to According to by ericfontainejazz
Thanks for the tests. However they are not mergeable as it since they fail :) Do you plan to change the code to make them pass too?
I don't see a real use case for 27 and 28 since some measures are completely ignored. Do I miss something?
In reply to Thanks for the tests. However by [DELETED] 5
Attached is modified 28 which has an added DC at end, making all measures reachable, which would be a use case in a hypothetical complex roadmap song (pretend that are measure 1 & 3 are actually long sections that you don't want to waste paper/ink copying...I have actually ran into these). I tried to make my tests as simple as possible while still exercising a specific case that might present an error.
UPDATE: Attached is modified 27 for "Jump to forward, Play until end" that has all measures reachable, which would be useful in a hypothetical case that you want 3 different beginnings to the same section (again assume that each measure may be much longer than it is and you want to save space).
I'll wait to see if someone else responds by saying something along the lines of these test cases are not useful. I won't be able to make them pass anytime soon, but I'll leave them up here...you may reject the pull request (I didn't know the tests had to pass!).