First off, just to be clear, this is a purely historical/theoretical question. In the example I'm going to give, I think it would universally be considered clearer just to write it in 4/4, maybe with accents, but I would like to know if there is a modern or historical notation for this kind of construction.

Say that I have a piece of music in 4/4, but at one point in the piece a measure is extended by a quarter note. The following measure is shortened to compensate, so that the overall rhythm fits neatly into its 4/4 surroundings. The effect is that it sounds like the first measure has gone on for too long, so that the second one has to rush to make up for it. I am only providing a contrived example here but assume that the piece justifies this.

Two bars of 4/4, with the first and sixth notes accented.

This is almost certainly the clearest way to notate this. Alternatively, the time signature could actually change:

A bar of 5/4 followed by a bar of 3/4.

This emphasizes that the 5+3 grouping is integral to the structure of the piece, but it removes the rubato-like push-and-pull feeling of the first notation.

I feel there might be some clever way to use a dotted barline to get this point across but it escapes me. This is the best I can come up with:

Two bars of 4/4. The first and sixth quarter notes are accented, and a dotted barline sits after the fifth note.

The problem with this is that it gives the impression of a 4+1+3 grouping, although the accents alleviate that somewhat. Making the actual barline dotted seems very confusing, although it would show that it should not feel like the end of the measure.

Finally, the time signature can be changed to 8/4:

A measure of 8/4 with a dotted barline after the fifth note.

This clearly shows the structure, but it gives the impression that all eight notes belong to one musical unit, whereas they're supposed to be separate measures. The best way I can think of to fix that is to separate the two in the time signature:

A measure of (5+3)/4, with a dotted barline after the fifth note.

Depending on the style of music, this is... opinionated at best, but it carries the meaning fairly well. The only problem is that it removes the "borrowing" or "stealing" feeling I'm going for.

Is there a notation, whether it is widely or currently used or not, which fully conveys the semantics of what I am describing? Of course I realize this is extreme pedantry and some of these examples are 99% there. But to be clear, here is the exact semantic meaning I am going for:

Two bars of 4/4. The first bar has five quarter notes. The second bar has a small "IOU" note, followed by three quarter notes.

However, I think this notation would take some explaining.

  • 4
    Do you plan to involve a conductor? Try to think what arm movements you want the conductor to do. The time signatures play a huge role in that.
    – T Andersen
    Jan 26, 2023 at 10:45
  • Related question and answer: music.stackexchange.com/questions/126910/… Jan 26, 2023 at 18:33
  • 1
    If it's to be for a small chamber ensemble, one option is to write pretty much whatever you want and specify that all players should read from a full score. Example: George Crumb, "Black Angels".
    – David
    Jan 27, 2023 at 2:26

5 Answers 5


If possible please avoid uncommon notation! It will often not be easier to read.

Maybe do something like this

enter image description here

This is more or less how Schumann does this in Mondnacht:

enter image description here


Simplest is to keep it in 4/4. The emphasis in bars of 4/4 is basically the 1st beat carries more. So, make the 2nd crotchet, 2nd bar, have an accent over it. You could also use phrase marks, if appropriate, so the 1st phrase is 5 beats long, the 2nd, 3.


I'll list my main points in headline form before expanding them:

  1. Meter does not always imply strong emphasis
  2. The highest goal should be clear communication
  3. There are no "rules," just conventions

So, #1: A time signature, on its own, will not necessarily tell the musician to put a big emphasis on a certain note. Yes, we often describe meters in terms of "strong beats and weak beats," and yes, downbeats often get an emphasis. And yes, some genres, like a Sousa march or a Viennese waltz, assume a pattern of strong metric stress ("ONE two THREE four," "OOM-pah-pah"). But it's just crass, and/or boring, to keep on slapping every downbeat equally. Even in these genres, phrasing means that some "strong beats" are stronger than others. And in a lot more music, our teachers yell at us to make "long lines," not to chop the music up with lots of stresses. If anybody played the Moonlight Sonata like a Strauss waltz I'd have to suppress an urge to slap them. So if you're contemplating this metric change just to get the performer to place an emphasis, it won't do the job on its own.

Instead, time signatures are more about how we think about the metric structure. The listener may or may not hear these groupings—or they may hear groupings based on emphases even if you don't change time signature—but changing it might help the performer think about it the same way you're thinking about it.

The first time I heard the "Dance of the young maidens" from Stravinsky's The Rite of Spring, I thought it must change time signature a lot. I heard a string of equal-length notes that I perceived as groups based on emphases: a group of 9 notes, then 2, then 5, then 3, then 4. Turns out it's notated just as plain-vanilla 2/4, with accents sprinkled in:

enter image description here

Here, the point is actually about fighting the time signature: most of these accents are on "weak beats"; their subversion needs a regular grid to subvert.

Point #2: If you do choose to change time signature, pick whichever method will be most easily understood by the performer, with the least explanation. Assuming that that's your goal, unless you want to make some mysterious inscrutable score as a philosophic art piece; I could imagine somebody doing that. But if the goal is communication between composer and performer, then pick methods that are conventional and simple.

If this grouping happens just in this one spot, I say go with your first example; just use a couple of garden-variety time signatures. If it happens all over the place, but in a regularly recurring way, then the "5+3" time signature is not at all uncommon. If the meter is kind of fluid throughout and doesn't really follow a fixed pattern, composers sometimes just drop dotted barlines here and there, or have no barlines at all.

Which leads to #3: You don't have to have barlines or time signature, or even a meter, or even a regular pulse. But many of these ideas have been already handled, and things like dotted bars or no time signature have been done. Inventing something brand-new might be interesting to make a point, but it would have to come with some textual explanations.

  • Re. #1: IME, people do expect barlines to follow the metre (syncopation notwithstanding). For example, I wrote a choral piece with an irregular metre. The first time we did it, I notated it in 4/4 (as I thought that would be easier for the other singers, and it sort of matched up some of the time) — but they seemed to have real trouble following it, and it didn't ’feel’ right. Next time we did it, I removed the time signature completely, and put barlines to fit the metre (giving bars of 3, 2, 3, 3, 2, 3, 4, 3, 4, 3, 3, 3, 4, 4…) — and the other singers seemed to ‘get it’ much easier.
    – gidds
    Jan 26, 2023 at 18:16
  • @gidds Yeah, I guess that's my point—they're more about affecting how the performer thinks about it than affecting what the listener hears. Jan 26, 2023 at 18:19
  • Good point about Stravinsky. His work came to mind while reading the question, and his solution makes more sense than the alternatives. Not having barlines makes it almost impossible for an ensemble to play together, certainly to rehearse bits!
    – Tim
    Jan 27, 2023 at 11:10
  • @gidds Your anecdote highlights the fact that the "numerator" of a time signature is often redundant since the reader can count how many beats are in each measure. I suppose its not entirely redundant in compound time, but all that need be indicated there is the presence of a prime factor of 3. If clarity were not the goal your piece could be notated as (3±1)/4.
    – Theodore
    Jan 27, 2023 at 20:15

If I'm understanding you correctly, I would consider a solution with multiple aspects:

First, I agree with Tim in keeping the example in 4/4. This way you show that 4/4 is the underlying conceptual meter throughout.

But using a phrasing slur in combination with this could help give the impression that, as you say, "the first measure has gone on for too long." The musician will clearly see five beats within that phrasing slur (and potentially another phrasing slur immediately following with only three beats) and get the impression of how things group together, their asymmetry, and the need for the second measure to "make up for it."

Another possibility would be to use your final example but without the "IOU" marking. This isn't completely unlike Messiaen's "added-value" technique where he inserted extra time into particular measures. You could simply inaugurate a type of "stolen-value" system where the additional beat(s) is/are stolen from succeeding measures. But of course you'd want an explanation specifying this in the music. The only issue I see with this approach is that conducting would be unclear; you'd probably want to clarify how these meters should be conducted, in which case you'd want to use a regular time signature.


I wouldn't change the 4/4 bar structure, but rather perceive and handle it as an accentuation pattern, as known from the drumset.

Drummers use it to spice up their rhythms. You can regard it as a special form of polyrhythm:

  • one voice runs on regular 4/4
  • the accentuated voice runs e.g. on 3/4, 5/4 or 7/4

The page below is taken from "The Language of Drumming", Benny Greb, Hudson Music 2012.

You can read the 3 applications at least in two ways:

  • as a comlete pattern, i.e. the 3/4, 5/4, 7/4 accentuations just stop once the 4/4 measure is played
  • as an incomplete pattern, i.e. you'd need to play a few more to "return to the beginning" of the pattern (i.e. to start with the starting-accent at "1" again), with the accentuation pattern shifting to the left or right from bar to bar

Hint: The notation shows the accents by hits to a certain drum. But you could also play e.g. quarter notes (or sixteenths as in the notation) at a moderate volume, while hitting harder (playing louder) for the accents.

Accentuation, Benny Greb

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.