bug: connected notes do not export to MusicXML
Hi, Developers, this is an isuue and bug report,
[Tech]
Windows XP SP3
MuseScore v1.2 revision 5470
[Issue]
Imported a MusicXML file from Finale, a piece in 9/8 rhythm with the notes connected with those horizontal bars (are they called 'beamed notes' in English?) in a 2-2-2-3 format, wherever possible. That looks perfect after import.
Then, after changing some style and page information, I stored the score again in MusicXML format with another name. When I reload this new XML, all the notes appear to be connected ('beamed') in a 3-3-3 style, which is absolutely garbage for a piece in 9/8.
It seems that MuseScore automates the connection of notes way too harsh. When saving a score, you should save what you have created, and open it exactly as you saved it. The automation of note-connection could be useful only when creating a score, but any change to note-connections should thereafter never be overruled when saving or loading a score to/from XML. This is a serious issue for anyone using odd time-signatures.
[Bug]
Created a new time signature: 7+11 / 8. Created a new score using this time signature. Saved the score in MusicXMLformat. When I try to load this XML, MuseScore crashes with a Windows notification 'program problem... need to close'. I was using a previously saved 'style', but unfortunately I can't say where the problem is, in the style, in the time signature, or anywhere.
But I am absolutely confident that MuseScore will develop into a program you can't ignore ! Go on please !
flyingdm
Comments
"which is absolutely garbage for a piece in 9/8."
Whilst I agree that MuseScore shouldn't alter the beaming, assuming that that has been stored in the MusicXML file, I can't agree that beaming in 3 groups of 3 quavers is a non-sequitur in 9/8, in fact I would definitely assert that it is the norm.
9/8 is the compound time equivalent of 3/4, having 3 dotted crotchets to the bar, each split into 3 quavers.
Any other interpretation of that is non-standard, and should be regarded as syncopation, which, I hasten to add, is perfectly acceptable, and is often the norm in some folk rhythms.
But, just to drive the point home, conventional music theory splits 9/8 into 3 dotted crotchet beats to the bar.
In reply to "which is absolutely garbage by ChurchOrganist
3+3+3 is definitely the norm, but there is one very famous piece that uses 2+2+2+3 - Dave Brubeck's "Blue Rondo a la Turk". I predict a 90% possibility that this was the piece in question :-).. And yes, it *should* have worked, and I was able to reproduce the problem with 1.2. But FWIW, when I took the XML file created in 1.2 and loaded it into 2.0, it looked correct (2+2+2+3). So it appears the bug was on the import side and has already been fixed.
As for the other issue involving the custom time signature, posting an actual sample MSCZ file and specific steps to reproduce would help. And if it turns out to still be reproducible with 2.0, then you should probably submit an actual issue to the tracker.
Beams should be imported correctly and crashes on import should not occur. Based on your description alone it is hard to pinpoint the cause of these problems. Could you attach the offending files for further investigation ?
Hello all,
Thanks for your answers ! I haven't found version 2.0 yet, is it already downloadable ?
As for the 9/8 (which is not Brubeck's), if it rhythmically WERE 3+3+3, you should i.m.h.o. normalise the score to be a 3/8. The point with (for example) a 9/8 is, that you need a way to know in a quick glance if the cadence is 2-2-2-3 or 2-2-3-2 or 2-3-2-2 or 3-2-2-2. These all exist, welcome to Balkan music!
Now the score that made MuseScore crash, is a piece in 18/8, or call it 7/8 +11/8. Just image that the notes were beamed in 6 groups of 3... you would have no clue to play it with the right cadence and accents, and you would have a hell of a job learning such music and remember it, because the melody line usually supports the cadence, and if you have the cadence not right, you won't be able to 'feel' the music as 'natural'. Even if you insert text "3-2-2+2-2-3-2-2" to explain the cadence, the score will be very difficult to 'read' while playing, if the beaming were wrong.
I must admit that our western system of music notation is not always suitable for complex rhythms/dances from the Balkans. Therefore it's difficult to discuss what is the 'norm', but good beaming helps a lot.
Leon, I attached the MSCZ and the XML that makes MuseScore v1.2 crash. The srore is in a beginning state, just a few lines.
Greetings,
flyingdm
In reply to 3-3-3 is not 2-2-2-3 by flyingdm
Ah! I wondered if you were referring to Eastern European music - the other guess was Latin folk music.
In that genre, cross rhythms of the kind you are referring to are common place, although I am by no means an expert on Balkan music, I did meet a Bosnian bass player a few years ago who gave me a little insight into the genre :)
Regarding MuseScore 2.0, there is currently no stable version to download, but you can preview one of the Nightly Builds here:-
http://musescore.org/en/handbook/comparison-stable-prerelease-and-night…
If you would like to try it out, download it, but please bear in mind that it is inherently unstable and not suitable for working on anything important - it may crash without warning for no apparent reason, taking your last hours work with it! However it is useable provided precautions are taken not to lose your work. You may also find that a score created in one version of the Nightlies will not load into a more recent version, so care also needs to be taken with this.
HTH
Michael
In reply to Ah! I wondered if you were by ChurchOrganist
Michael, since I am playing percussion, these are my first steps into music notation whatsoever (I can't even read notes), so at this moment my scores are not 'important'. However, I would like to have means to share written music in my little orchestra, so it will become important, and eventually I will learn to read notes? The others can read notes of course...
On the Balkans, there is different understanding of rhythm, due to dances that belong to it. A dancer learns steps, where a long step has a duration of 1.5 times that of a short step - very simply said, of course there is much more. So, any odd time signature could be (in our system) described as pieces of 2 and 3 notes. For example, a signature of 11/16, is often cadenced as 2-2-3-2-2, but there are also dances in 3-2-2-2-2.
Did you know that an 8/8 signature must be an odd rhythm? That's because it is cadenced as 3-3-2 or even 3-2-3. If it had NOT this odd cadence, but instead a regular cadence like 2-2-2-2, you would have to describe the music in your notation as 4/4, filling a measure with 1/8' notes though. One should choose the simplest time-signature for a piece of music that is possible, regardless of the length of individual notes in it. A 4/4 measure can contain op to 32 1/32' notes, but you wouldn't describe that music as a 32/32 measure, would you? Interesting, how people interpret these things differently. I know only one piece of music from the Balkans in 9/16 signature that is interpretet as 3-3-3, and only at the very end of that song. So, in this case only the melody-line with very clearly 9 notes belonging together determines the time-signature as 9/16, otherwise the signature would be 3/16 (having more measures).
I wouldn't speak of cross-rhythms here, just odd rhythms. On the Balkans, you'll hardly find syncopics (go to Africa for that), but especially the Greeks are capable of elongating a measure on the feel, without breaking the basic rhythm... just putting a few microseconds extra to a signature - a feeling quite hard to learn in for example an extrordinarily slow 9/2 or 5/2 measure.
And back to what it was about... we definitely need the beaming for Balkan music, as 25/16 and 31/16 signatures just EXIST out there...
Greetings,
flyingdm
In reply to Ah! I wondered if you were by ChurchOrganist
By the way, fellows, we are in the 'bug report' thread, and my problems are likely to be solved in a couple of weeks. Thanks to all devopers !
If anyone likes more discussion on time signatures, I will be glad to receive private mails, see my account details, tabpage 'contact'.
flyingdm
In reply to 3-3-3 is not 2-2-2-3 by flyingdm
> if it rhythmically WERE 3+3+3, you should i.m.h.o. normalise the score to be a 3/8.
FYI, that might be the norm in Balkan music, but not in most Western music. 6/8, 9/8, and 12/8 are referred to as "compound meters", and are always written in groups of 3. That is, 6/8 is always 3+3, 9/8 is always 3+3+3, and 12/8 is always 3+3+3+3. Again, I am talking about most Western music, not necessarily the Balkan traditions in particular. In fact, if you wish to write 9/8 written any way other than 3+3+3, you would normally be expected to write that into your time signature - calling it 2+2+2+3/8 rather than 9/8, because 9/8 pretty universally means 3+3+3 in the Western tradition.
But indeed, regardless of time signature, beaming overrides should be honored in MusicXML export/import, and apparently it isn't in some cases. Leon, here is the 9/8 file I created as a test. Load this into 1.2 and export as MusicXML, then import that MusicXML into 1.2 again, and you'll see the beaming change. Load that same MusicXML into 2.0 and beaming is preserved. So I assume the bug is on the import side and has already been fixed, but you might want to verify and maybe add this as a test case.
In reply to > if it rhythmically WERE by Marc Sabatella
Hello Marc,
For sure overridden beams should be preserved on export and import. I will have to install a 'nightly' build, and test your file, since v2.0 is not available yet. I will look into Brubeck's music anyway on YouTube, I don't know him :-))
Even if I would create and use a time-signature like 2+2+2+3/8, then MuseScore should adapt the beamings accordingly to that, wherever possible (haven't seen or tested this yet...) The VERY best for MuseScore though, would be to let me assign just a 9/8 (or any other simply expressed signature, as is practice in balkan music), and give me an option to design the default beamings (cadence) like for example 2-2-2-3 for the whole piece, and also the possibility to override this for some system or for a particular measure, because also that exists in balkan music: just one measure cadenced differently. THAT would make MuseScore the best in town ! Overriding beams is already possible, but with a bit more work, though.
A related problem... I can design a 2+2+2+3/8 (=9) measure in MuseScore, but NOT a 2+2+3+2+2/8 (=11) measure, because there are only 4 fields available for compound rhythms. It would therefore be better to set no limits here: let the user assign a 31/16 time signature at will (already possible), and let him describe a beaming or cadence of (imaginary example) 3-2-2-3-2-2-3-3-2-2-2-2-3, or even 7-7-10-7 at will, and give an errormessage if that does not fit to the assigned time signature.
I am glad with so much response to my report! That makes me confident that MuseScore will develop into an allround tool.
Have a nice evening,
flyingdm