Stopping Playback Note Sustaining Bug

• Oct 16, 2010 - 00:55

MuseScore Version: 0.9.6.3
MuseScore Revision: 3507
Operating System: Windows XP
Sometimes when I am stopping a piece of music manually (by pressing the play button while it is playing), some notes, that I find to be in the middle of when I stop the playback, hold their note for infinite amounts of time. The only way to stop it is to find the note that is making the interference and click the note. This sometimes happens to multiple instruments. Please tell me if there is a way to fix this, and try to improve this on the next release of MuseScore.

-imazi643

Comments

• Oct 16, 2010 - 03:40

Yes, I've found this too. Not all the time, but it has happened.

The complete workaround (if clicking the note doesn't work) is playing the score (or at least the section) again.

Reply

• Oct 14, 2011 - 19:20

FWIW, I generally use TimGM6mb.sf2 or the modified version, which is as small a soundfont file as I have seen, and it makes little difference on my system (which I have described plenty enough on this thread, but it is AMD and generally runs 6-9% CPU when Musescore is playing music.) The # of times a note sustains when stopping playback seems to be about the same as with a larger or smaller soundfont.

Reply

• Oct 15, 2011 - 00:10

This is a known MIDI problem.

You will find that all the high end MIDI sequencers and also apps like Finale and Sibelius have a "Panic" button to deal with this.

It is caused by MIDI Note Off events not being received by the Sound Module or SoundFont.

There is a MIDI command for sending a global note off on all channels, which is what the "Panic" button does.

With today's faster processors it doesn't happen as often as it used to :)

Muse Score should have a similar panic button to deal with it on the rare occasions that it happens - if you like I will dig around in my MIDI documentation for the relevant MIDI message - if I can remember where I stored it - lol

Reply

• Oct 16, 2011 - 03:09

This isn't so rare on the 3 systems I have been working with. And the level of processor doesn't seem to make that much difference either.

I wonder why the developers of Musescore (and Sibelius, and Finale) don't execute a global off (twice) when playback is stopped.

Reply

• Dec 20, 2011 - 21:16

I'm not sure that sending Sequencer Stop message twice would help.

Once this phenomenon occurs you can send as many sequencer messages as you like and they will be ignored - that's my experience of it from wayback anyhow.

I'm pretty certain that this is a problem between the FluidSynth MIDI engine and some processors - I have never experienced this on any of my systems, which are pretty low end btw :)

You might find the other thread on this subject helpful.

http://musescore.org/en/node/11873

Reply

• Sep 4, 2011 - 17:12

I have separately reported this for 1.1.

I provided one file that demonstrated it, but it was a complex file. I then provided a second cleaner file, but I don't think anyone was paying attention by that time.

It does not appear to be related to CPU speed, soundfont, or the particular computer I am using. I have seen it happen on Windows XP, and 2 different Windows 7 computers with plenty of RAM and CPU power, using several different soundfonts, including the one that ships with the product.

Reply

• Sep 4, 2011 - 21:57

As I mentioned elsewhere, I have had this happen with 3 different computer systems, one running WinXP and 2 running Win7. One item that is the same between these computers is that they all use AMD CPU's.

It generally happens once a working session or so, when I enter music and frequently playback the music to check my work. Maybe once an hour during a working session. Maybe a little more often.

Reply

• Sep 12, 2011 - 04:20

The file you provided, plays fine for me. As mentioned early note sustaining issues are related to the speed of the computer. As I understand it, If the computer is too busy at the exact moment the note-off signal is sent then it may not get acknowledged. The same problem can happen in any notation software that uses MIDI for playback. Finale has a special menu item to specifically address this issue. See http://www.finalemusic.com/usermanuals/finale2011win/content/finale/Pla…

Reply

• Sep 12, 2011 - 04:59

>> The file you provided, plays fine for me.

I don't think this is file related at all. I have hundreds of files that I work with and any of them will exhibit the problem if I work with it long enough.

>> As I understand it, If the computer is too busy at the exact moment the note-off signal is sent then it may not get acknowledged.

That is interesting information and I can see how that could happen given how MIDI works. But wouldn't it then also happen when playing the music normally, without the user stopping playback? I believe the note-off signal must eventually happen with each note that is played. The "MIDI lock" problem only happens to me when I stop playback.

Also I do wonder if performance is the whole story. My system does not appear to have difficulty handling Musescore's load. CPU usage running Musescore generally runs at or below the 5%-7% level and while running Musescore I have not seen it exceed 15%. (I realize that is just a sampling of CPU usage and it could spike higher). The system I generally use has 6Gb RAM and generally runs with over 4Gb RAM free, so I don't think virtual memory is the culprit. My system uses an AMD Athlon II X4 630 which, while not the hottest CPU available, I suspect is more than up to the task. In other words, it happens far too often for my to buy into the performance explanation as the entire answer.

The fact that I have seen this on several different AMD based computers makes me wonder if there isn't a connection there.

>> The same problem can happen in any notation software that uses MIDI for playback.

FYI, I have various versions of Sibelius and Finale, and other MIDI based software, fairly extensively, and I have never experienced this with those products. I see it regularly with Musescore.

Reply

• Sep 12, 2011 - 05:51

I know that on my old laptop it would happen, but I don't remember it happening in MuseScore since I purchase a new laptop a year ago (unless I happen to be compiling software in the background). I don't remember the specs of my old laptop but it was AMD (low-end, seven years old), my current laptop uses Intel i7.

FYI, the MIDI lock problem has been happening every couple of hours for me in Finale 2007, not during playback but during keyboard note entry. (I'm using Finale a lot more than MuseScore at the moment because of the requirements of the projects I'm working on).

Reply

• Oct 9, 2011 - 05:57

The most significant problems for me are outlined on the Milestones page. The "Basic Notation" section is probably the most significant, since the issues apply to almost every score I want to create. Although I can workaround many of the problems, the manual changes take extra time and are error prone.

For scores with a lot of text, I come up against bad kerning on printouts (plus Mac OS screens) . It's fixed on the trunk but the trunk is not stable enough to use.

Reply

• Dec 20, 2011 - 20:56

"As I understand it, If the computer is too busy at the exact moment the note-off signal is sent then it may not get acknowledged. "

This may be the case, for this happens mainly in a typically busy section.
But, this still happens in some cases when it may even be a single instrument playing, and the error will follow.

Reply

• Jun 12, 2012 - 17:57

I don't know if anyone will see this, but just to be complete...

Re: "But wouldn't it then also happen when playing the music normally, without the user stopping playback? I believe the note-off signal must eventually happen with each note that is played. The "MIDI lock" problem only happens to me when I stop playback."

Recently I had the note sustain problem show up, one time, during playback without stopping the music playback. Win 7 system with 6Gb RAM, AMD processor, Merlin_Vienna soundfont, CPU not under any particularly heavy load, running only MS 1.2 and whatever else Win7 was running at the time.

Reply

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