Merge Voices together so they are chords in a single voice?

• Nov 16, 2014 - 17:52

I posted a previous topic at http://musescore.org/en/node/38726 trying to copy individual staffs into one. I have now done that following the suggestions given by Jojo-Schmitz (thank you again!) but now the notes are in four individual voices on the piano staff.

Is there any way to "flatten" or merge the voices together so that all of the notes turn into Voice 1? It's really messy the way it is right now, with note stems all over the place and identical notes where two or more voices were singing in unison being shown multiple times. Since the score is 16 pages long I don't want to have to manually delete duplicate notes each tine they occur and I can't see any way to fix the stems either.

See attached screen shots. In "First system - black" you can see how it looks when I print it. In "First system - coloured voices" you can see how each note in the chord is attached to a different voice and for example the first chord in the treble clef has green and red both showing a D, so that should only be there once, and the first beat in the bass clef is a mess with two rests and a note. I can manually delete the rests because they are not on voice #1, but the second chord is so messy and hard to read. The stems for the three notes that are the same length should go up and the stem for the blue one that is a different length should go down. I think this would all happen automatically if all four notes were in the same voice as a chord rather than individual voices sharing a single staff, but I may be mistaken.


Comments

After fighting with it for 20 minutes I've managed to fix the first two beats of the song.

It involved insisting to the software that I really wanted to be editing voice #2 (it kept going back to #1) then holding shift and typing the note name I wanted, and then putting it into the right octave with the mouse manually after when it went into the wrong octave.

There must be a better way.

This is how much I've managed to fix now (screen shot "Measure 1 - coloured notes.PNG") after 20 minutes, and "Measure 1 - coloured notes - fully fixed.PNG" after another 10-15 minutes. Very frustrating.

I now realize that "voices" are the way to go when I want to have notes starting at the same time with different lengths, so measure #1 ends up having three separate voices used due to the first two notes being a whole note and dotted half note respectively. So I guess it couldn't flatten the entire thing into a single voice, but if all places where multiple voices are playing same-length notes at the same time could be collapsed into one voice (automatically) that would be so much better than the hoops I've had to go through to get this to happen.

Recognizing, of course, that this is the first time I've used this program so it's quite possible I'm just doing things the worst possible way :)

There is a plugin called "implode" that does the job more automatically, at least in the easy cases. You might try that instead. See the Plugins menu at right of this page, look under Explode. However, it's limited in a number of ways. In addition to requiring the same rhythms, it doesn't handle ties, and I think tuplets don't work well either if I recall correctly. There is also the

Note for 2.0, as of a few minutes ago when the change got merged (!), there will be a native Implode facility that does more or less the same thing but less limited. Still, it will only work well if all parts have similar rhythms.

In reply to by mjzwick

The implode function (both the plugin for 1.3 and the new native facility for 2.0) does not combine *voices* of a single staff; it combines *staves*. So you would use it if you have each line on a separate staff. I think there are separate plugins that can help with combining and or spliiting vocies within a staff.

In reply to by Marc Sabatella

Closest combination I can find is Voicecaster to break into a score with separate staves, then implode, then copy back to the original score. This seems pretty clumsy when all you want to do is combine two voices in the same staff that have the same rhythms. I brought the starting point for this score over from ScoreCloud, which has a simple click and drag feature for managing voices (however, it has other deficiencies that had me export the score to Musescore via MusicXML). Using 2.0 Nightly Build (for the feature set) and this is the only thing currently inferior in Musescore. Hand jamming all of these notes is going to be painful!

In reply to by mjzwick

I'm curious how you found yourself in this situation in the first place. I guess maybe something about ScoreCloud encourages you to write in separate voices even though you know full well that's not what you want?

Anyhow, I was going to say maybe someone could write a plugn to do what you want more directly, but then I had a thought. Right now, the new 2.0 implode facility does nothing if you have only a single staff selected. It turns out it is simple to tweak the code so that if you only have a single staff, it does a voice-wise implode (and delete the notes it implodes, so you are left with a nice clean single-voice staff. So after a little more testing, I think I'll submit that for consideration.

I also started to wonder what the analogue might be for exploding a staff with notes in multiple voices, then it occured to me - that just a matter of first imploding into a single voice then doing a regular explode.

In reply to by Marc Sabatella

@Marc

"I also started to wonder what the analogue might be for exploding a staff with notes in multiple voices, then it occured to me - that just a matter of first imploding into a single voice then doing a regular explode"

whilst tapping my head and rubbing my tummy. ...

I'm so glad you know what you are talking about ! LOL

In reply to by Marc Sabatella

ScoreCloud does let you work in multiple voices within and across staffs/staves. I suspect that these notes, while tied into chords, were in different voices. In ScoreCloud, had a noticed it, I could have gone into the Voice viewer and just dragged them together. The Voice/Mixer view in ScoreCloud provides shading and coloring to see where all of the voices are, which can be combined, split, and rearranged quite easily. I never used the voice capabilities much in Musescore because I didn't have much need and because using voices in the score was somewhat clumsy - it was easier to just use two staffs/staves.

I think I will leave these sections while I work on the rest of the score. If you do get around to posting something into the nightly build please let me know and this score can be a test guinea pig.

Thanks!

In reply to by mjzwick

The implode voice facility I described is now in the builds. Again, it isn't really the general purpose convert-anything-you-like-between-voices-and-staves facility that the discussion morphed into. That is currently doable by use of copy/paste, exhcange voice command, and perhaps the selection filter could be involved too. The new facility is just meant for the very specific case where your prats have the exact same rhythm - the same very specific case the original implode facility was intended for.

The way it works is, if you select >1 staff, then the notes from voice 1 of each staff are combined into chords in voice 1 of the top top staff - just as is already the case. but if you select just one staff, then the same baasic thing happens for the voices of that staff - the contents of the tfour voices get collected into chords in the first voice. The notes in the other voices are then deleted, since you presumably don't need them any more.

Again, this is meant for the specific case where you have identical rhythms between parts - just as the explode faciltiy is. A totally separate facility for combining arbitrary staves into voices within a stave or the reverse would still make a nice project for the future. But since it's a less common use case, and perfectly possible to do already using the tools provided (copy/poaste, exchange voice, selection filter), I can't see doing anything on that any time soon.

In reply to by Marc Sabatella

YES :-)

I've been thinking about this sort of thing for a while. Here's maybe a little more food for thought.

If I'm arranging for a big band with, say, 4 trumpets, 4 trombones and 5 saxes, it'd be nice to be able to have a condensed score with all 4 trumpets on one staff in separate voices, The 4 trombones in 1 staff in separate voices, 2 altos together in separate voices, 2 tenors, and a bari. So, I've dropped my score size for the instruments from 13 staves down to 5, and it's easier to see the voicings in the instrument groups.

Now, I'd like to be able to toggle this to a full score with all 13 staves, and also I'd like to be able to have the parts extractable from the voices on the staves.

It sounds like the functionality you're talking about submitting might be able to accomplish this sort of thing quite well.

Thanks again for all your great work Marc!!!!

- Mike.

In reply to by carneyweb

This would be awesome IF the voices in the staves that become parts stay linked so that I can edit the score and those changes would be reflected in the extracted parts!

The score that started this thread is Tpt 1&2, Trombone, Alto, Tenor, Bari, Piano, Bass and Guitar. The Trumpet and alto/bari parts could be scored together in voices on the score, thus reducing the number of staves by 2. Some of my music has two trombone parts. A big band, as Mike says, would benefit even more.

Thanks!

In reply to by carneyweb

You're welcome :-), but I think you'd find the implode/explode feature too limited for this. It really is by its very nature limited to having exactly the same rhythms in all parts. I mean, it hoepfully won't crash if you don't, but it wouldn't preserve all information, either. So it couldn't be used that way except in the very simplest of cases. And even then, there is no facility to do *exactly* what you are describing. The combining of staves into separate voices you'd still need to do manually, by first doing voice exhcange to put them into separate voices, then copy & paste to combine onto one staff. My implode change would allow that to then be easily converted to a single vocie - again, if the rhythms all match - and you could then explode it again. But it seems a pretty indirect way of going about what you describe.

I think what you are proposing would really need to be separate - a way to combine the contents of up to four staves into voices of a single staves, completely independent of the rhythms involved, and also the reverse. That's perfectly possible to implement some day as well, but really a totally separate thing.

In reply to by Jojo-Schmitz

Yes, as I have said, the underlying assumption of both the explode and implode facility is that all parts have the same rhythm. All bets are off if you violate that. For implode, we are simply adding notes to the existing destination staff/voice - it's rhythm wins. Trying to handle any other cases makes things much much harder, and in any case, is simply impossible in the general case.

In reply to by Jojo-Schmitz

No, I think that's worse. Now you have an "orphan" note in voice 2 - a note on beat one with nothing on the other beats. That's bad. Makes it very hard to edit that measure later. You and I might know the workaround (exchange voices, then back again) but the avergae user won't. I think it important a built in function not create these "holes".

In reply to by Jojo-Schmitz

Merging voices with different rhtyhms loses information *by defintion*. It's just a matter of whether you lose information and create and measure that is then very difficult to fix by hand, or lose information but at least make it easy to go in by hand and decide how you want to fix it.

The only alternative I see is to abort the entire operation and restore the original state if you enter any cases of mismatched rhythms, But that seems unnecessary. The facility is useful enough as is; why limit it arbitrarily?

In reply to by Marc Sabatella

merging voices my way doesn't lose anything. It just merges what can get merged and leaves the rest as it was. And easy to fix too, just, in the above example, delete the quarter and add back as 2 tied eighth. This non-merged chord is also quite easy to spot visually, as opposed to spotting a lost note.
Another option might be to skip the entire measure in case of a rhythm mismatch?
Also we may special-case multi stave implode (or restrict my method to single staff implode), as there no information gets lost.

In reply to by Jojo-Schmitz

Merging your way doesn't lsoe information because you don't actually merge at all. This might suit your case, but the purpose of implode was to produce a single staff/voice, and your change defeats that, plus it creates an essentially corrupt measure. It might be easy to fix in this particular case, but in general, measures with orhpan notes / holes are impossible to edit fully unless you know the special magic voice exchange trick.

It just seems like you are wanting implode to be something very different from what it was originally intended to be. The whole purpose of the feature was to produce a *single* voice/staff out of multiple parts. And by definition, that requires identical rhythms.

Now, it is true that I did not originally intend to handle multiple voices at all. So I guess I can't be too opposed to a change in how that particular aspect works. But I am very uncomfortable creating what are essentially corrupt measures, even if a privileged few will know the special workaround for fixing the corruption.

In reply to by Jojo-Schmitz

As a compromise: what if we suppressed the merge for unequal durations, and kept any "extra" notes* in voices 2-4, but in any measure where this is an issue, we replace the merged notes with (invisible) rests rather than delete them? That way there are no holes. And if you don't mind holes, you are then free to delete the rests.

*By "extra" notes, I mean, what if the *soprano* has the quarter, alto has the two eighths? We are currently merging the first alto eighth with the soprano quarter and deleting the second eighth. So at least what results looks sensible and is clearly singable. Your change would keep the first alto alto eighth but still delete the second, which I would suggest is *worse* than turning merging the first alto eighth into with the soprano quarter. Now the alto has literally nothing to do on the "and" of the beat.

And and BTW, when I say "what if we...", I really meant, "what if you..." :-)

In reply to by Marc Sabatella

OK, so let us ;-) have a look at that
Actually I don't think there is any problem with trailing rests missing, or is it?
So it is only leading and intermediate rests we'd need to deal with?

And yes, now I see that my current fix loses data too, quarter in voice 1, 8th in voice 2, implode and every 2nd 8th is gone, bad....

In reply to by Jojo-Schmitz

some examples generated from the corresponding mtest:
before:
a.png
after, using the current code:
b.png
result: the 2nd quarter note from voice 3 is lost, the first turned into a half.

Some alternatives (manually generated):
alternative 1, keeping notes that can't get merged, keep rest, visible or invisible:
c.png
alternative 2, without (the trailing) rest (and here my PR currently fails too, it loses the 2nd quarter of voice 3, updated PR should fix that):
d.png
alternative 3, rewriting voice 1 using tied notes:
e.png

In reply to by Jojo-Schmitz

I think the expected result depends in part on what one's use case is for this feature. The one that strikes me right now based on what you are saying is, a chorale-type piece (SATB, mostly homophonic in terms of rhythm) where you wish to create a piano reduction. That means combining voices where conveninent, but not to the extent of losing information. In this use case, it is not so crucially important that it be possible to follow each part individually, so it's OK if a part moves between being a separate voice one one beat but a note within a chord in another voice on the next beat, and there is no need to show the rest on latter beat.

With this type of use case in mind, either of your first two alternatives do the job. In the first case, I'd expect the rest to be made invisible. In the second, we have a "hole", but you'll be glad to hear I'm coming around on this point. The problem of "holes" is a general one that we need a better solution for, so I'm going to assume that at some point we'll have one. So I withdraw my objection. I think #2 is the way to go. It looks good, and it's way easier than #1 or #3.

BTW, I was remiss in not commending you on taking the initiative here! Sorry if it seemed I was being harsh. I could be wrong, but this might be your first "featured" related PR? Nice job!

In reply to by Marc Sabatella

Yes, indeed it is about SATB scores (as these are the usual scores I'm working on).
And #2 is what my PR is doing now. And yes, I'm glad to hear that you're now fine with those 'holes', as I still have no idea to get that fixed ;-)
But no, I didn't regard your remarks as being to harsh, it was a good discussion IMHO and worth being discussed.

In reply to by Jojo-Schmitz

Glad to hear! In my head, I can see how to avoid the holes by replacing the merged chords with rests and adding them to a list and setting a flag if we encounter any unmerged chords then checking the list of rests at the end of each measure and removing them all if there were no unmerged chords. Soemthing like that anyhow. But I now say, skip it, let's live with the holes and hope for a better way of dealing with those measures someday. Aside from that, I do rather like the results.

In reply to by Marc Sabatella

>> I'm curious how you found yourself in this situation in the first place

Don't know about the OP, but in my case it came out of the PDF converter that way. I was baffled by the extra rests scattered around the score til I figured out that the PDF converter arbitrarily put some of the notes in a second voice, and inserted rests into the melody line. The score is actually just a simple monotonic melody (hey, I'm a banjo player, not a musician!).

The implode function deleted the notes that were mistakenly placed in the second voice, and preserved the rests, exactly the opposite of what I was trying to do. These rests are very tenacious, I cannot delete nor replace them !

I tried converting to XML, and tinkering, but that way madness lies ...

Still struggling with getting this sorted ... I may just color the rests white and get on with my life ...

In reply to by BanjoJake

If you post the score you are having problems with, we may be able to help. Just attach it to a reply using the "File attachments" link right below where you type,

If it was truly a monophonic line, then there are never notes at the same, only notes in one voice and rests in the other. In which case, it may work well to select all (ctrl+a) then press the Voice 1 button in the main toolbar to try to move everything to voice 1. This will leave rests behind in voice 2, but those can be delete by using the Select Filterer to exclude voice 1, select all, and delete (and then reset the filter).

Also, rather than coloring this white, it is usually better and much simpler to mark them invisible by pressing "V" or using the Inspector. They show greyed out on screen but won't show at all in the actual score.

Depending on what we see in your score we may have other suggestions.

In reply to by Marc Sabatella

Hi, Marc, thank you so very much for the quick response !

>> it may work well to select all (ctrl+a) then press the Voice 1 button
>> in the main toolbar to try to move everything to voice 1.

Unfortunately, this did not work. Here is the file, straight from the PDF converter web page.
R23B_2015-07-23_202533.PDF_.mscz

Please note the problem in measures 32 and 40 (at the moment I'm concerned only with 'Whiskey Before Breakfast' - the tune, not the activity 8?).

In case it's interesting, here is the file I uploaded, from a scanner:
R23B (2015-07-23 202533).PDF
It was originally rejected, until I noticed that the music had been printed at a slight angle on the page.

All the other problems, like the codas, I cleaned up, but figured it would be better to send you the raw files.

In reply to by BanjoJake

Are you using the current version of MuseScore (2.0.2)? I tried the select all, voice 1 trick, and it worked perfectly for measure 40 and partially for measure 32. Most of the the notes from voice 2 moved to voice 1, replacing the rests that were in voice 1 and leaving behind only rests in voice 2, which are then easily deleted. The only note that didn't get moved was the first note of measure 32. This didn't get moved because it is a quarter note, and there is only an eighth rest in voice 1. The line that appears to start on beat 2 actually is starting on the "&" of 1. So, first move it half a beat later, then do the select all / voice 1 trick, and it will be fine.

To move those notes later, select them (click first eighth note in the measure, shift click last) then ctrl+x to cut. You'll then want to turn the initial eighth rest into a quarter rest by pressing 5, then click the rest on beat 2 and ctrl+v to paste.

In reply to by Marc Sabatella

Hi, Marc, thanks for the response, I've tried several times now, on two different PCs, and I'm not seeing any change in measure 32. I am using v2.0.2, and measure 40 converts perfectly, but measure 32 does not change when I do select all, voice 1 !
It remains as in the original:
voice 1: an eighth rest, 6 eighth notes and an eighth rest
voice 2: a quarter note

Anyway, I'm slowly learning how to move notes to where I want them using cut & paste & arranging the rests ... like chess moves ...

In reply to by BanjoJake

Yes, you need to move the notes first, as I said - otherwise, there just isn't room in voice 1 to move the quarter note from voice 2. Currently, there is only an eighth rest. I meant, you could simply lengthen the eighth rest - click it and press 5 - but all the notes in that measure are half a beat too early.

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