I'm part of a small group and we have our own pieces, which I'm not allowed to share. But I'll describe the situation and hopefully get the point across. In nearly all the songs in my group's repertoire, different instruments have repeated bits at different points, which means that usually there can be no volta brackets or other repeat signs.

However, we often need to have single-staff scores at hand so that we can actually read the music while rehearsing, otherwise it's impossible due to the sheer size it gets. That said, I can either:

  1. write the entire piece without a single repeat, explicitly showing each phrase again and again; or

  2. use volta brackets and segni etc. in different places, relying heavily on the fact that each score you look at contains a single staff for a single instrument.

For example:

When you're writing the flute part all by itself in a dedicated sheet of paper, there's no reason you couldn't put volta brackets around measures 11 through 20 rather than writing them twice, even if the other instruments do not repeat said measures. Right? No need to waste many pages on music that has already been written, read and understood. So let's say I choose to go with that. Now after the flutist plays measures 11 through 20 the second time, they would actually land on measure 31 rather than 21, big picture-wise. Think communication with the other musicians, especially considering they all have different repeats themselves. Also somebody watching the rehearsal and counting the measures would have counted 30 of them so far.

If I don't change the numbers in this way: “Let's play again from measure 21. I mean, it's 21 for me, but it should be 31 for somebody who doesn't have any repeats on their score up to this point. It should be measure 18 for the violinist. No, wait, that is… 17, I think…?” vs. If I do: “Let's play again from measure 31, everybody.”

So in the end what I did is, I put a little 31 over the ending volta bracket (the one that closes measure 20), effectively changing the number of the following measure, which would otherwise be 21. I tweaked it and counted from 31 on, afterwards. By folding measures 21 through 30 into the repeat since they are 11 through 20 played again, I deliberately removed those numbers from the flute score.

A more visual example would be the following picture: Each rectangle represents an entirely different score, although all scores here belong to the same piece.

Note that the numbers are not actually printed every measure; this is just the rationale.

Is it something that should be avoided? If a pro saw this, would they feel inclined to undo everything and put the numbers as they were, leading to the need to remove the repeats altogether? I'm feeling a little insecure as I don't want my group to have poorly-organized sheet music for posterity.

  • If you don't want to leave "poorly-organized sheet music for posterity", the simplest plan would be to create a full score without any of these repeats. Any notation program will have a cut-and-paste option, so it's easy to do this. Any future problems understanding your parts can then be solved by referring to the full score.
    – user19146
    Aug 2, 2016 at 13:25

7 Answers 7


As I understand bar numbers, they are intended to locate the correct position in the score, not the point in time. Note, that they often do not exist at all, and you will find rehearsal marks (squared or circled characters or numbers). I never encountered the "repetition-reflecting" numbering you propose, which has a set of drawbacks on its own: in your example "B" the player will have a hard time to arrive at bar 6, espcially if more than one bar is repeated (single bar repetition would anyway be typically represented by a bar repetition looking similar as ".//.").

So a conductor will typically state something like "We start with bar five, second repetition".

So summarized, perhaps your numbering is not wrong, but so unusual, that I would tend to avoid the necessary explanation for justifying it.

  • Good point on the repeated bars - // - which may also have 'by the way, there's 7 of them' somewhere.
    – Tim
    Aug 2, 2016 at 7:46
  • Thank you for replying. The problem with "bar 5, 2nd repetition" is that bar 5 would be in a different place in every staff score when using numbers traditionally. That is, if A has a repetition of which bar 5 is a part while B does not, then the number 5 will be in different timestamps for them. A could be playing the 8th bar of the song as a consequence of looking for bar #5 on the page. I need to humbly clarify that my group is far less invested in sheet music than I am. They won't even bother trying to read if everybody has different bar numbers at any given time due to misaligned repeats. Aug 2, 2016 at 8:43
  • @user3241796: I guess, if you have differing numbers of repetitions depending on the voice you could consider to switch to use excercise marks instead (and possibly keep the bar numbers in standard way).
    – guidot
    Aug 2, 2016 at 9:22
  • Thank you. I'll look into exercise marks. Judging by the context I have a hunch they work a little like a Segno—just without the Dal Segno command written in the score. Aug 2, 2016 at 9:34
  • 1
    @guidot the usual English term for "exercise marks" is "rehearsal marks".
    – user19146
    Aug 2, 2016 at 13:15

I agree with guidot, but a side discussion started in the comments that I wanted to address in a top-level reply.

user3241796 says:

The problem with "bar 5, 2nd repetition" is that bar 5 would be in a different place in every staff score when using numbers traditionally. That is, if A has a repetition of which bar 5 is a part while B does not, then the number 5 will be in different timestamps for them.

In making the parts, you should have each part be structurally identical in almost all cases. That is, you shouldn't have one part have a repeat and another not.

I can imagine an exception: you might have a piece of music with long repeated sections, where only a subset of the ensemble does something different. Maybe there's a very long "verse" with the same accompaniment but with a different melody/solo on top of it. If the length were enough to take most of a page, then it might make practical sense (from a typesetting perspective) to write the accompanying instruments with a repeat, but write out both repetitions for the melody instruments. In this case, I would recommend writing both sets of measure numbers in the accompaniment, as suggested in the question.

  • 2
    I agree with the guidelines you proposed, but I'm afraid it has fallen to me to transcribe other people's pieces, pieces that either have never been written or have been written only as an orchestra sheet music, where misaligned repetitions evidently need to be unfolded. It all comes down to preserving works that do not preserve the traditional conventions of Western music. We call the solos 'ouchi' and the acompaniments 'jiuchi'. We're a taiko ensemble, actually. Jiuchi is usually a constant sequence of a super-short loop, like augmented eigth + sixteenth (meaning a single-beat loop in 4/4). Aug 2, 2016 at 15:37
  • 1
    This didn't fit in a single comment. I meant to add that, due to the overwhelming length of several kinds of jiuchi, there is no way not to fold the repeats into volta brackets when writing a single part score for a single part musician to read. Just imagine most of a page repeating that example above — augmented eighth + sixteenth. don ko don ko don ko don ko don ko don ko don ko don ko.......... Just writing this here in plain text feels like too much repetition of such short a phrase. See why I need to break down the structure of different parts at different times? There's no escaping this. Aug 2, 2016 at 15:41

It sounds as if, in this case, writing 'true' bar numbers would be best. Where a bar within a repeated section has two bar numbers, indicate them as e.g. 12(24).

You must not be using a computer scoring program. This sort of independent looping is tricky to achieve, without breaking the connection between Full Score and Parts, with all the potential for error that encourages!

And only consider giving different parts different 'road maps' when there's a substantial reason, e'g' a jazz number where the rhythm section repeat the same 32-bar chorus several times, with varying solo and background arrangements in the front line. Don't do it capriciously.

  • I use Lilypond. I think I'll be able to tweak the bar number engraver with some research to include both numbers. Thank you for the suggestion. Aug 3, 2016 at 0:13

As far as I remember, Guitar Techniques used nearly the same notation that you propose when they used to do songs. They had the bar number at start of each row, for example 47 51 55. And then, if the last row had a repeat, next row might start with bar 63 or 61.

I wasn't even aware some are numbering the bars as they appear on sheet not as they appear in song. "Fifth bar, second repetition" sounds quite wrong as that just depends on how the sheets are done not how the song goes. At least for a normal band it would seem quite ridiculous that the bass player would have sheets and sheets of the same notes just because the singer's part has some slight changes from verse to verse.


At each 'new' part, maybe 16 or 24 bars in, where everyone's together again, or a different theme starts, use rehearsal letters, which would be the same for everyone. Then, if you said 'let's take it from 4 bars after C',or '6 bars before J', all would start in the same place.

  • Thank you for replying. I like your idea of labeling measures with something other than numbers so as not to override them. Truth is though, I'm not sure that would work out well because the structures of our pieces are packed with repeats, which puts a lot of those 'start 4 bars after C' occurrences inside repeats. At any point we often need to start at the 17th repetition of A's short phrase, in the middle of B's solo. We use 'performance-oriented' labels so much and 'page-oriented' numbers so little that I think we'd be better off without the traditional measure numbering system altogether. Aug 2, 2016 at 8:31
  • Most solos are not that long. I can't understand why you'd want to start in the middle of a solo. It hardly is in context if that's the case.
    – Tim
    Aug 2, 2016 at 8:34
  • I'm just writing songs we've learned across the years. Songs from masters that visited us to host a workshop, songs some of us composed without sheet music, etc. This group isn't classically trained, so this kind of thing tends to happen a lot in compositions. A lot of melodies with no pinpointable key due to lack of a tonic, etc. You said most solos aren't that long, and you're right. But in my group one person is often stuck in a loop for ages before another person begins their solo. That's why I planned on shifting the focus of the bar labels from the written music to the performed music. Aug 2, 2016 at 8:51
  • 1
    It sounds as bar numbering may not be very appropriate or useful in your type of music. Label sections instead.
    – Laurence
    Mar 26, 2017 at 12:09

Notation is not right or wrong but good or bad. Its function is to convey meaning to musicians. The question is how good that notation is for that purpose, and how efficient.

With complex repeat structures (dal segno al coda), repeats may be traversed several times, requiring a stack of numbers.

You save redundancy for things like "let's start in measure 47 in the second repeat" but the question is whether this redundancy arrives at players with multiple numbers, and your principal goal is that every player may have individual repeat structures, so a decision like "ok, lets keep the repetitions in the da capo" is not possible to make.

Even without complex repeat structures, you'll need to put bar numbers (and where applicable, their ghost stacks) at the start of every repeat and at every alternate ending end at the exit.

There is a reason that generally, repeat structures are common to the whole orchestra. Particularly for drummers, there are abbreviations like bar repeats (that look like percent signs with a variable number of oblique bars in the middle) that can significantly condense repetitive material without affecting the overall repeat structure or bar numbering.

So while at surface value, your proposal may have some merit, it does come with complications particularly of having problems scaling up. There may be some cases where it appears worth the trouble, particularly in lesson books or other stuff where introducing your own conventions and benefitting from the savings covers a larger body of material under your control and you can evaluate the drawbacks and benefits specifically.

As a general practice for orchestral scores, having the same repeat (and bar numbering) structure for everyone in a manner where no extra bar numbers are required at start and end of repeats is very much the norm.


At least according to the Royal Conservatory of Music, obeying repeat signs is optional. This is supported by sources such as the 2022 RCM Piano Syllabus (https://rcmusic-kentico-cdn.s3.amazonaws.com/rcm/media/main/about%20us/rcm%20publishing/piano-syllabus-2022-edition.pdf) and https://www.olivermusicstudios.ca/RCM-Practical-Exam-Info---Piano-Levels-8---9. In fact, the 2022 RCM Piano Syllabus says this in its Page 4:

Repeat signs should ordinarily be ignored.

Repeat signs are indeed still not always followed in public and published performances of music. For example, this sheet music excerpt from the Breitkopf und Härtel edition of Beethoven's Appassionata Sonata shows the first repeat ending, for the development and recapitulation, of that sonata's 3rd movement: Piano Sonata No. 23 in F Minor, Op. 57, Movement 3, First Repeat Ending, Ludwig van Beethoven, Breitkopf und Härtel Edition

These two performances of the 3rd movement of Beethoven's Appassionata Sonata flagrantly skip that first repeat:

For another example, these sheet music excerpts of Mozart's Piano Sonata No. 16 in C Major, K.545 show the ending repeat sign, for the development and recapitulation, of that sonata's 1st movement:

Breitkopf und Härtel Piano Sonata No. 16 in C Major, K.545, Movement 1, Ending Repeat, Wolfgang Amadeus Mozart, Breitkopf und Härtel Edition

Schirmer Piano Sonata No. 16 in C Major, K.545, Movement 1, Ending Repeat, Wolfgang Amadeus Mozart, Schirmer Edition

These two performances flagrantly skip the repeat of the development and recapitulation of that Mozart piano sonata first movement:

The thread https://classicalmusicguide.com/viewtopic.php?t=15240 brings up even more examples of skipped repeats in published performances, including of Haydn symphony movements.

This jump measure numbering that assumes that repeats will always be followed therefore cannot be supported.

  • 1
    I don't understand the way in which you're using the RCM instructions. The RCM (and ABRSM) requests students ignore repeats when taking exams because it saves time. It has nothing to do with how the music would typically be performed. A performer not following the repeats — i.e., not following the composer's instructions — would be making an exception to the expected practice, and so also an exception to the measure numbering.
    – Aaron
    Jul 15, 2022 at 6:06
  • 1
    These guide lines are for a single performer taking an exam. They don't really apply to OPs situation. Jul 15, 2022 at 6:50
  • @Aaron - I have heard "exceptions to the expected practice" of following repeats often enough that I do not think the expected practice actually involves always following the notated repeats. Take performances of the 3rd movement of Beethoven's Appassionata Sonata, for example - these performances skip the first repeat of the development and recapitulation entirely: youtube.com/watch?v=LqIKv-42v1E, youtube.com/watch?v=xnqP6a9YDf0
    – Dekkadeci
    Jul 15, 2022 at 17:48
  • @Dekkadeci In that regard we agree, but I don't think relying on the RCM exam procedure is enough to make your case. Adding your previous comment to your answer, plus another example or two, would strengthen your post considerably (especially if you could point to a numbered score of those pieces and show what the editorial practice is).
    – Aaron
    Jul 15, 2022 at 17:52
  • @Aaron - Added the comment above to my answer, then - it didn't take long - found another example of skipped repeats in performances, this time of Mozart.
    – Dekkadeci
    Jul 15, 2022 at 19:47

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.