Combining beams from different voices

• Mar 28, 2021 - 18:03

I recall reading a couple of months ago that Musescore was adding an option to combine beaming for voices 1&3 and also an option for 2&4. Did this ever get implemented? I can't find it in the preferences or manual.


?? You mean a note from voice 1 beaming to a note from voice 3? That never had been implemented and also doesn't make any sense.
You can slur and even tie different voices though (even it esp. the latter doesn't really make any sense either)

In reply to by Jojo-Schmitz

In makes sense in 3-voice guitar tab pieces because it allows an independent harmony in voice 3 whilst combining the beams to give the overall rhythm and keeping it correct for the time signature. Guitar Pro has this option.

The volume of voice 3 is then easy to adjust relative to voice 1 but without a proliferation of beams.

In reply to by Marc Sabatella

I won't be able to find any in published music because there's no computer playback to cater for and therefore the human player would naturally sustain notes correctly. This is only about computer playback of scores to keep the rhythm clear, the note durations correct and allow easy volume assignment by voice.

Here's an example of the same measure shown three times:

1) Three voices to keep the note durations correct for playback but the rhythm is confusing

2) Two voices to keep the rhythm clear but added ties to keep the note playback duration

3) Note durations as first measure and stems/beams merged for clarity.


Variations of hiding ties, stems can help but it's a lot of manual effort compared to rhythm merging and is a neat feature in Guitar Pro.

In reply to by frfancha

For playback.

It's not a notation that only I would understand, it is used in Guitar Pro which is quite a well known and respected product but maybe it's a bit too niche a feature for a general score editor to support.

A published score, whether as individual sheets or in a book, does not have to support playback so it's inherently not a issue.

Guitar Pro music published in their online Songbook supports this notation.

In reply to by yonah_ag

It is not possible to write what you want in a regular staff. Maybe it can be done by combining voices on a computer, but again it doesn't make sense in terms of note-writing, printing, reading.

Since I play the guitar instrument a little bit, I understand what you want to do.
The only remedy for this is to issue the "let-ring" directive to the player.

In this way: the player will provide continuation (hold) in possible strings.

About playing the score on the computer: I think you can get a moderate result (ie enough to give an idea), although not exact.


Attachment Size
z-example-1.png 11.88 KB

In reply to by Ziya Mete Demircan

It's possible and is supported in Guitar Pro. I have made various workarounds in MS using either 2 or 3 voices but it's labour intensive. A human player would not need any further instruction, it is only to make the playback sound better and it makes a huge difference i.e. reduced "choppiness".

Another guitar tab system, "TablEdit" recognises this situation and has a 1-click function for doubling the playback duration beyond that actually notated.

In reply to by yonah_ag

Yes I know.
I know that in time (when I was an arranger) I manually adjusted the guitar notes one by one in the midi-file. And this process was no different than digging a well with a needle.

Soon, I realized that the time wasted by trying to imitate the player one-on-one was not worth the time lost compared to the quickness of getting the job done the easier way. (Even better, Using a real guitar-player would be more effortless and time-saving.)

I also know that the "let-ring" we currently use in the Musescore software is no different from the "Sustain pedal".
A software dedicated to the guitar may have done this further. But this feature is not yet fully available in Musescore.
And I hope someday there will be a contributor who will define "let-ring" to work according to the string information. And when that time comes: the "let-ring's" we write now can magically come to life.

I think the way we should head is not to abuse the voices. Of course, there is no rule, person or law that prevents this. With the software in his hand, whoever does what he wants, the way he wants. But I think that if we do a process in the right way, Also it will work correctly in the future.

In reply to by yonah_ag

Excellent comments yonah_ag and Ziya Mete Demircan!

This is a really important request and could help MuseScore become wildly popular among strings musicians and those using tablature notation. I hope development is listening with an open mind to the possibilities raised here ... and to other recent guitar and tablature requests.

As I understand your request here yonah_ag:

You'd like to use simple rhythmic notation when writing for guitar AND you'd like score playback to sound as if played on guitar. The sustain pedal is a likely candidate BUT it can easily sustain too much, as Ziya Mete Demircan pointed out. And he rightly commented that tediously selecting notes in a Piano Roll editor to realize proper guitar-sounding durations is indeed nothing but tedium—and it's simply not a tenable solution on long pieces.

I've started a new topic on this and attached a sample score that illustrates part of the issue.. Hope to see you there!

I'm very interested in any progress on the plug-in!


In reply to by scorster

Yes, this is what I would like to achieve. The best work around is multivoice but then it really requires stem/beam rhythm combining, ("abuse"), Guitar Pro style. Incidentally, GP allows any of the 4 stave voices to be above or below the staff and performs visual rhythm combining accordingly, e.g. you could have voices 1, 2 and 3 above with 4 below.

The MS sustain pedal is really designed for piano and it therefore sustains all notes in a measure right thru to the end of the measure. This is not always possible because of subsequent notes on strings which need to stop the sustain.

The MS Let Ring line is really just a sustain pedal applied to a single note and therefore has the same problem of sustaining right thru another note. Another issue is that sustain stops with the end of a measure. So a note played on the final beat of a measure cannot ring.

My plugin is very manual. It allows me to set the duration of selected notes to any value up to 2000, (up to 32,000 would be better but this is capped by the plugin API). It's a quick means of applying PRE changes. It would be nice to develop some automation in it. (The plugin also gives me fret hand colour coding of tab numbers).

In reply to by yonah_ag

If I understand correctly, you are suggesting abusing voices just to simulate the natural rining of guitar notes, where even in a single voice, a note plucked on one beat will naturally still sustain into the next, thus overlapping with it?

This is not multiple voices, and using the incorrect notation to simulate this playback effect is the wrong way to go. You're talking about a playback effect only. So either find a soundfont with an appropriate long sustain "release" (e.g., the sound doesn't stop immediately when the note off message is received), or edit the playback info (e.g., in the piano roll editor, or using a plugin). These the more logically correct ways to do it, and also it would doubtless be easier to actually do in your score, plus it wouldn't mess up MusicXML import/export, wouldn't hopelessly confused blind musicians trying to understand your score via keyboard & screenreader, etc.

In reply to by Marc Sabatella

It's only 'abusing' the rhythm info not the musician playback or computer playback. In fact a guitar player would have no trouble reading the rhythm, it would simply look like 2 voice scoring.

The handbook says, "A voice is a musical line or part which can have its own rhythm independently of other voices on the same staff. Voices are sometimes called layers in other notation software."

In this situation I have 3 lines with their own independent rhythm, (and volume), on the same staff. Isn't this multiple voices? This is more than just natural sustain, I have independent rhythms. I also use a sound with an extended ring which does help. If you extend the ring too much then the playback sounds awful - like a long droning.

I have a plugin which sustains the notes but it's still quite manual. The easiest way to do it is Guitar Pro style automatic rhythm combining.

The MusicXML could be preserved if the rhythm combining was only visual but retained the actual voice 1 and 3 in the score file. I'll see what GP does with this.

In reply to by yonah_ag

I would be happy to see something like a plugin etc to automaticly go from more voices to a tablature where the rhythm is condensed to one line of rhythm-signs (voice?). That is exactly what lute tablature intabulations makes out of pieces with (or for) more voices

In reply to by wolfgan

That is essentially what I would like MS to do but it looks like a non-starter.

Currently my plugin works from condensed rhythm, (i.e. a single scored voice covering 2 guitar voices), and allows extended duration of selected notes. It's also manual at the moment but a lot quicker than using the PRE.

Using multiple voices but merging the rhythm would be preferable because I would then have separate volume control at the voice level. I'm not sure that it's possible this way with a plugin because it would just be a display convention rather than any change to the underlying music.

In reply to by yonah_ag

It is abusing the feature in that it using something intended to produce a particular notation effect and trying to coerce it into producing a desired playback effect that has nothing whatsoever to do with the actual purpose of the feature. It's the wrong solution. You don't have a notation problem - the normal notation is perfect as is. You have a playback problem - the correct notation doesn't produce the desired effect in playback. So the solution should come on the playback side.

If you have something with independent rhyth, then yes, that's multiple voices, but by definition, "independent" means "don't beam them together*. Together is the opposite of independent. So keep the independent parts independent, sure, but within each each part, it should be be a single voice, and if the playback doesn't ring the way you want currently, that's where any new feature should come.

In reply to by wolfgan

A beautiful example Wolfgan, both compositionally and as a example of the issues discussed here. (What is the piece/composer?)

I tool the liberty of adapting your example score and have submitted here for review and discussion.

• I changed the meter to 2/4 to make the metric situation clear to all.

• I changed the tablature to a "guitaristic" presentation with fret numbers on the tablature lines.

• I slowed the tempo to make the issue more obvious to the listener.

    1._Learning_tablature Interpreting durations.mscz

It’s clear that notational and publishing standards exist and work quite adequately in most "fretboard" circumstances. These notational norms work well enough because guitarists know they need to interpret durations.
• There are norms that lean on multiple voices to attempt to show all durations, or to emphasize the critical durations.
• There are norms for simplied rhythms whick I call “temporal onset” rhythms.

That’s the beauty of standards. There's so many to choose from!


Some closing thoughts:

• The inherent complexity of notating and interpreting guitar music cannot be fully expressed strictly by relying on separate voices. And sadly those efforts, though extremely valuable, usually fails to provide "fretboard accurate" playback via MIDI.

• Over simplifying (writing a continuous stream of eighths) omits the option of notionally showing critical durations. But it's easy. And it naturally produces "temporal onset" rhythms which can be difficult to achieve with voices employed.

• I’m glad to see the recommendation downplayed that we might achieve the desired "fretboard" durations by using a particular soundFont.

Hopefully we’re getting somewhere on discussing issues that could make Musescore extremely attractive to those notating for fretted instruments, with the goals of:

• Unvoiced notation that accurately sounds "Fretboard-accurate Laissez Vibrer"
• Voiced notation that accurately sounds "Fretboard-accurate Laissez Vibrer"
• Multi-voiced notation that can alternately show "temporal onset" rhythms in treble clef, Tablature, or both!


In reply to by yonah_ag

In order to answer I'd need to know which of the various examples posted here you are referring to, what the correct published notation for that is, and which specific note or notes you want to sound different from what the default would be. But in general, correct notation means, everything that logically / musically belongs in voice 1 goes in voice 1, everything that logically / musically belongs in voice 2 belongs in voice 2, etc. No concern for how MsueScore might happen to play it, I mean, to create correct notation that other musicians would understand properly and would not get an engraver fired for putting into a published edition. Then from there, again, if some particular note you want to overlap with another in the same voice because of the specifics of how guitar sounds work, that's something you tweak via the playback facility, which will hopefully be enhanced to make this easier in the future.

In reply to by Ziya Mete Demircan


Looking at the first half measure in scenarios (1) and (2) : All the info is in the .mcsz but not all visible in the displayed or printed TAB.

(1) Is written in 2 voices and the upper voice notes are quavers. A human player would naturally let these ring for their intended crotchet length but a computer would play them as quavers and give a 'choppy' sound. (Depending on the tempo this effect could be mitigated with a particular soundfont - but you cannot assume that users will all have the soundfont).

(2) is written in 3 voices and the upper voices notes are now crotchets so that computer playback is correct. Human playback will also be correct. However, the rhythm stems/beams have been combined or 'abused'. You can't tell from these images because I have removed the colour differentiation.

So even a savvy musician would not detect this 'abuse' by visual inspection. This could be implemented as a visual device without affecting the underlying score and thereby not confusing screen readers or accessibility.

This visual device is used in Guitar Pro when there is more than 1 voice above (or below) the stave.

In reply to by yonah_ag

In fact the underlying score only needs to be kept for things like MusicXML export. Screen readers would simply see the layout as (1) above.

This seems to be one of the cleanest ways of displaying 3 or 4 voice TAB and would not impact any linked standard notation. There is also no need to resort to playback tweaks to get the correct note duration.

TAB has always been a bit 'looser' than standard notation and I doubt there is even a defined standard for this situation. Perhaps GP, as a tab specialist program, provides a sort of de facto standard.

In reply to by yonah_ag

Indeed, the human musician will have no idea what limitations the computer software imposed on the engraver. But if you saying that this is the correct desired notation, and that you wish it to sound different in MuseScore than notated, once again, the solution isn't to hack MsueScore to do something musical meaningless like beam notes in different voices. It's simply to allow more playback control. just because Guitar Pro refused to provide this simple playback control and instead resorted to this hackery, that doesn't make it good. It's not, it's the wrong solution.

If Guitar Pro refused to provide a hammer but instead handed you a screwdriver and told you to use that to pound a nail, that doesn't mean we should too. We should simply provide a hammer.

In reply to by Marc Sabatella

On the contrary, GP provides excellent playback control as well as mixed beaming. It is MS which requires hackery in the form of PRE. There is no easy way in MS to keep this simple measure tabbed correctly and playing back correctly without hacking the underlying MIDI durations.

GP does give you a give you a hammer, (in the form of a hammer-on), and a whole load more effects in its toolbox: pull off, palm mute, tapping, brushes, slides, bends, vibrato, tremolo bar, rasguedo and more all 'out of the box', no MS style workaround hacks needed. This is not surprising since GP obviously specialises in guitar. (Fortunately there is no screwdriver and the only nails anywhere in sight are fingernails).

MS wins out on sheer controllability of layout and probably supports many instrument playback features pretty well but nothing like a program specialising in a few instruments.

The beaming is not meaningless: any TAB reader would have no trouble understanding the meaning without needing to know anything about voices, e.g. on a printed sheet.

Is there any reference guide to TAB that says mixed voice beaming is invalid? I fully accept that you will be right regarding the incorrectness of this type of beaming in standard notation but I only see it being useful in TAB.

In reply to by yonah_ag

If someone can post one single example of a correctly notatied score that you believe does not provide correct playback, and cannot easily be made to provide correct playback by simply lengthening the playback of the notes you feel are incorrect. please do so. Should be a single post with one single example and requiring no more than a couple of sentences of explanation - one to say which notes you feel aren't being played properly, another to say why you believe it wouldn't be fixed by simply lengthening the playback duration of that note.

If such a score exists, then inventing brand new notation could be considered as a last resort workaround. So far I've yet to see any evidence of a need for this, though.

In reply to by Marc Sabatella

It's not brand new notation in the world of TAB and I am not suggesting that it has any place in standard notation.

By lengthening the playback do you mean only by changing the note type, (crotchet, quaver, minim, dotted etc.), or does this include using the PRE? If you can clarify this point I'll post a short example.

In reply to by yonah_ag

Vocies are voices whether talking about tab or standard notation. If you prefer to post a pbulished example in tab format, that's fine too, just do that and also the corresponding MuseScore file representing the normal correct way to enter it, then describe what specifically you don't like about the playback. Then we can discuss how to take the correct notation and alter the playback to be more what you would like. I can't tell you what alterations that would be until you provide the published example and MSCZ file, though.

In reply to by Marc Sabatella

if you take a look at the published example from Mar 30, 2021 - 10:23 you can easily see what I mean:
the Lute tablature in line 1 is the only correct way to notate the music from the grandstaff or the guitar notation from line 2+3. All the beams + rhythms of all voices have to be combined in one line above. Ideally Musescore should automaticly change the notation of line 3 (or even better to allow one tab staff to be linked withe a grandstaff) in a linked staff with lute tablature to what you see in line 1

In reply to by Marc Sabatella

Here's another published example:

1) Songbook.png


Viewed as TAB and Notation:

2) TAB_Notation.png

(Incidentally, this second measure is incorrect since fret 6 on string 5 cannot be a dotted minim due to the subsequent 2 on the same string. Also, the quavers in the middle voice of measure 1 would be better as crotchets for playback but the temporal onset beaming would be the same).

Hiding notation and setting rhythm as follows:

3) GP_Settings.png


4) 1-Above_2-Below.png

which is a clear representation of how the piece sounds rhythmically.

Changing the rhythm setttings to all above stave gives:

5) All_Above.png

which obscures the melody rhythm.

In reply to by yonah_ag

Sorry, I'm trying my best to understand but really am having trouble sorting through the very many apparently unrelated different examples and observations being made here and trying to reconcile them all to understand the special unusual case being discussed here. That's why I asked for one example - a single published excerpt presented as a single scanned graphic and accompanying MSCZ file representing the standard correct MuseScore equivalent, and a very simple two-sentence description listing:

1) which note you =is not playing back as you would prefer
2) why you believe you would not be able to get it by altering its duration

Everything else just gets in the way of my understanding - I'm overloaded with too much information as it is. That's why I'm asking for one simple example I don't have to spend hours trying to understand.

In reply to by yonah_ag

Yes, I think a new thread would be useful. One example that involves one very brief scanned excerpt (I can't see why more than one measure would be required), one MSCZ file showing how you've tried to enter it into MuseScore in a notationally-correct way, and then exactly two sentences, one saying which note is not playing with the duration you want, there second explaining why you feel changing its playback duration would somehow not do what you want either. Keep it that simple, and maybe I'll be able to follow!

In reply to by Marc Sabatella

BTW, when I say "published", I don't mean some random GuitarPro file uploaded to a score sharing website. I mean an actual piece of published sheet music from an actual music publisher. If files uploaded by random users who may or may not understand music notation are the only examples you can find, I'll still try my best to understand, but then the discussion is more likely to turn to, what would have been the correct way to notate it in order to get the desired result.

In reply to by Ziya Mete Demircan

Of course, that is standard notation, and I'm coming to understand the perceived issue is with playback of correctly-notated tablature only. And I can certainly see that using three voices as we see often here is not as practical for tab. And I know that one way this is often solved is by using a technique similar to drum notation, where we rely on the convenient fact that note duration isn't really something one directly controls much of the time to combine notes of different "logical" durations into a single chord (eg, four snare drum quarter notes simultaneous with eight eighth note in the ride cymbals can be represented in one voice by rotating the snare as eighths and combining them with the ride).

So I could certainly see that happening in this piece, someone making the conscious choice to tell that little lie, pretending a dotted half was an eighth in order to show it in the same voice as an arpeggio in the tab. In which case, simply increasing the playback duration would allow it to ring longer.

In reply to by Marc Sabatella

It is necessary to look at it from two angles:
1. Notation.
2. TAB's playing.

Writing as Standard Notation is an easy issue. There are many sources on this subject.

The subject of playback-interpretation (for guitar and similar string instruments): Each string should be considered as an independent voice (for standard guitar: 6-string / voice).
This problem is not seen in normal staff notation, but if you look at the TAB-staff (think as if there is a monophonic channel for each line / string) you can understand the problem.
And of course this is not a notation problem, it's just a playback interpretation problem.

Suppose the "let-ring" is written to Staff:

Normal Notation and TAB:

Colors: shows the length of time the let-ring is held.
Numbers: Indicates the moment the String was played.


In reply to by pabale4

I'm sorry you are feeling that way! One thing to keep in mind here: we are all just fellow users helping each other out. There is no official "team" of staff people helping "customers". Or more to the point, we're all on the same team here - the team of users. So, the best way to improve the average level of respect is simply to contribute respectfully wherever possible. Also, we all have our good days and not-so-good days, and we try to cut each other slack when a fellow MuseScore use who is volunteering their time to try to help happens to be having one of those not-so-good days.

Another thing that's important to realize is, English is not the native language for a lot of the users who volunteer their time helping here. That often has a lot to do with why responses can sometimes seem curt.

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