Bug report: corrupted import of a GuitarPro file

• Mar 1, 2017 - 06:31

Hello !

Using the nightly build of 2017-02-27, I tried to import a GuitarPro file and got quite a "funny" score ( see the picture).

I attach a zip file containing the guitarpro file, a pdf showing how it looks in GuitarPro, and a picture of show MuseScore imported it.

I am willing to work together with anyone working in this bug, to help troubleshoot ( I can for example generate test files in guitar pro, or something like that).

Although I cannot support my statement with reliable statistics, I believe that GuitarPro is a software that has a BIG market share among guitar players, so to have in MuseScore support for it might open many doors for Musescore (e.g. open the doors for users of GuitarPro who want to migrate to a more powerful notation software and carry along their GP scores).

Thanks

Ariel//

Attachment Size
crazy_import_result.zip 1.62 MB
crazy_import_result.jpg 568.76 KB

Comments

Hello,
Several different things:
1) Contrary to what you say, the score is not corrupt. What you see (in your attached image) is the result by importing the GP file into a recent version 3.0 dev. Eg:200889f

Consider that these versions are under development, are not really reliable, are intended primarily for testing, not for doing real work, and that we must wait a little so that all this stabilizes.
You are not observing these results by importing with 2.0.3 3c7a69d, or the current 2.1 dev., right?

2) Did you use the feature (Measure -> "Freely") or not in GP to enter your score? I do not know enough about this feature in GP for commenting on it. In any case, I note that by checking this feature "Freely" in GP, before export, the time signatures are stabilized after importing into MuseScore: 1test_GuitarPro_file.gpx

3) I do not know why, and do not understand why you entered your score, essentially melodic, with GP in Voice 2, or with a mix that seems incoherent between voice 1 and voice 2? Is there a particular reason?
1) test_GuitarPro_file.mscz

After exchange of voices, and some cleanup, you can receive: 2) test_GuitarPro_file.mscz
or if wished: 3) test_GuitarPro_file.mscz
The result would have been better by entering the score with a single voice (or essentially) about the display of the rests.
(it remains some cleanup to do for remove extra texts "let ring", "P.M. and "Barré", which are displayed in x copies. I had pointed this out already, but apparently this is not considered a problem requiring a fix)

Finally, I noticed the slurs in particular (glissandos too maybe) have not been imported. And I observe, from scratch, that it is the same in voice 1 or voice 2 for example. I'm a little surprised by that. I do not remember that problem.
I do not have time to check further today. I'll see soon.

In reply to by Nicolas

Hello lasconic !

I would gladly generate some GuitarPro test files, with their corresponding pdf:s, in order to help fix the current bugs/limitations. Just let me know exactly the description of the contents of the test files.

For example something like this:

"make a test file # 1 with guitar pro, containing 5 measures, with a barre in measure 2"

"make a test file # 2 with guitar pro, containing 5 measures, and a barre that starts in measure 2 and ends in measure 3"

etc,etc

Thanks

In reply to by ArielAr

I don't know how "let ring" or "barré" works in GP. Just create one file with 4 quarter notes and "let ring" on top of them and attach it to this issue https://musescore.org/en/node/30336

In general, one or two measure with the construct we want to test is already good to test and fix the problem.

For GP6 format specifically, it would be great if someone could work on the spreadsheet here https://musescore.org/en/node/98786 so we can have a better mapping between the GP and MuseScore instruments.

In reply to by Nicolas

Hello !

I did what you said for "let ring " and added test files to that thread.

I also created test files to troubleshoot
- "barre"
- ad libitum/rubatto/free time, which I include here.
- Palm Mute ( a pizzicatto-like effect of using your palm to quiet the strings)

Feel free to ask me for more test files.

And thanks for your commitment to MuseScore and great work !

In reply to by cadiz1

Hello Cadiz1 !

And thanks for your input. Let me answer your questions:

1) I want to help Musescore to become the best possible music software. I have not tried importing into older versions, if something is broken in the latest night build, I would gladly contribute to having it fixed. I do not expect any software to be bug free, I just want to help improve it.

2) Yes, this is a jazz piece that begins with a long rubbato/free time/adlib passage. In GuitarPro the measure bars become dotted and greyed. "Ad lib" passages, which in my opinion are very common in live performances of jazz music, are already straining for most musical software. I hope we can help the import functionality of Musescore to handle it as best as possible.

3) This is a guitar piece, where the bass and the melody have "different lives on their own". It is not a monophonic piece as in the case of a flute. I usually enter the bass in voice 1 and the melody in voice 2. Sometimes I have to resort to 3 voices. when they move independently, that is the nature of music in polyphonic instruments. In guitarPro you can enter even a piano piece so 4 voices can also happen.

I understand that it would have been easier if the music had only one single voice - one melody line, but I have higher expectations for Musescore than just being a software for handling simple, monophonic melodies :-)

if you want to debug/troubleshoot and you want me to generate different guitarpro files to test things then I would gladly help. My goal is just contributing to make MuseScore even greater than it is. Let me know how I can help.

And thanks for answering my thread.

In reply to by ArielAr

- "Sometimes I have to resort to 3 voices. when they move independently, that is the nature of music in polyphonic instruments. In guitarPro you can enter even a piano piece so 4 voices can also happen.
I understand that it would have been easier if the music had only one single voice - one melody line, but I have higher expectations for Musescore than just being a software for handling simple, monophonic melodies :-)"

Of course, MuseScore supports four voices with .gpx files (but since a year, in my memory, not more ...), and it's, in your file, a very secondary point.

- "I usually enter the bass in voice 1 and the melody in voice 2."

I think you will get better results by doing the opposite.

Sorry, I do not have no more time to answer any further. And my major concern by now is to check about import of slurs and so on.

In reply to by ArielAr

Please do not think of the current released version (2.0.3, or 2.0.3.1 for Mac) as "older versions". These are the *current* versions. The nightly builds are "current"; they are works in progress and not intended to be actually usable. They might or might not happen to not crash instantly or wipe out your hard disk, but you shouldn't be trying to actually use them for anything resembling real work. We certainly do welcome people *testing* things with them and reporting things that don't work, but please try to be clear that is *all* the nightly builds are intended for. And if something doesn't work in a nightly build, it will also be important to determine if it is is something that works in the released version or not.

And yes, definitely better to use voice 1 for melody, voice 2 for bass in MuseScore and most other software - the assumption is voice 1 is higher than voice 2. If you habve a score where this was done backwards, first thing you should do is use Edit / Voices to flip them.

In reply to by Marc Sabatella

Hi Marc !

I was never attempting to do "real work" with the nightly build. I just wanted to check how far MuseScore had come in importing GuitarPro files. Given that there are continuous improvements in MuseScore, it would be a pity to report a problem or missing feature in an official release that might have already been fixed since long in the nightly build branch.

You wrote "And yes, definitely better to use voice 1 for melody, voice 2 for bass in MuseScore and most other software - the assumption is voice 1 is higher than voice 2. If you have a score where this was done backwards, first thing you should do is use Edit / Voices to flip them."

I was tempted to answer in a private message, but given that the issue could be of general interest I chose answering publicly.

I thought it did not matter which voice had the bass. It is the first time I hear it.

5 minutes ago I saw this tutorial about multiple voice for GuitarPro and they place the melody in voice 2:

https://support.guitar-pro.com/hc/en-us/articles/200633951-Video-GP6-ed…

But I will check with GuitarPro support what they say.

I would like to ask more specifically what kind of problems will Musescore exhibit if the bass is in voice 1 instead of voice 2.

And what about the other voices ( 3&4 ) ? Is there also a preferred range for them ?

Thanks in advance !

Ariel//

In reply to by ArielAr

First, yes you are right that it is good if, upon finding a problem in the current release, you also check to see if it is already fixed in the nightly build before reporting it. I just wanted to make sure you understood the relationship because when you said something about "older versions", I was afraid maybe you were misunderstanding.

The reason it matters which voice is which is that the convention in music notation is to have the upper parts use stems up, the lower parts stems down, so the stems point away from each other & outside the staff. And MuseScore - like almost all other software - defaults to making voice 1 have stems up, voice 2 stems down. So putting the upper part in voice 1 allows everything to work the way standard notation expects.

I think the confusion comes in because in the tutorial you pointed to, the melody is *not* the upper part - it actually is somewhere in the middle. As you can see from the notation in that example, the melody entered in voice 2 has stems down. Looks fine in this case because voice 1 was not in fact a bass part - it was a chord with both bass and treble notes. But if it had literally been a bass part and a treble part, then in Guitar Pro as well as MuseScore you'd have wanted the treble part in voice 1 so it could have had stems up. Otherwise - again, in both programs - you'd have your stems pointing the wrong way.

So again, the *upper* part - regardless of whether it is melody or not - is supposed to have stems up, according to the universally agreed-upon rules of music notation. And the way to get that in both MuseScore and Guitar Pro is to put it in voice 1.

In the cases where *more* than two voices are needed, there is no specific musical expectation about which way the additional stems would point. It is not that common to need more than two voices, and in the cases where it *is* needed, the specifics tend to be different for each note and so it is hard to generalize. But MuseScore - again, like almost all other software - defaults to putting voice 3, voice 4 down. Meaning, if you do need additional voices and are wondering what voice to put them in, it usually makes sense to choose voice 3 if you want an upstem, voice 4 if you want a downstem. But only you can decide which way you want the stem to go. On the other hand, if it is *exactly* three voices consistently throughout a passage, but the inner voice sometimes would look best with upstems and other times with downstems, then I'd probably put it all in voice 3 because *logically* it really is just one voice, then flip stems manually where you want them pointing down for readability reasons. That's easier than putting some notes in 3 and some in 4.

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