Approaches to structure

Hi everyone,

I often take forever to complete pieces, so my mission lately has been to find ways of being more methodical with my writing. I don't mean paint-by-numbers but just to not reinvent the wheel every time I sit down to write. In some regard, the painfully slow pace stems from insecurity emotional and creative. But I think there are holes in my theoretical knowledge as well that contribute to a lot of false starts and backtracking.

Which brings me to musical structure. In terms of structure, I pretty much play it by ear and almost never am able to decide on a structure pre-compositionally, as much as I have tried. I'm wondering what other people's approaches to structure are and if anyone can recommend some helpful concepts and/or resources on this topic.

For context, my pieces are generally pretty conservative in terms of structure. Most of my pieces usually have an ABA, ABAB'CA, or ABACA structure. I like to keep the material unpredictable, so I'll often have some major deviation in style or texture and have a lot of connective material. It's not that the end result is unbalanced. It's just that it takes so long, because I'm going mostly by intuition.

Sorry if all this is pretty vague. I guess what I'm looking for are "handles," as in concepts that can help me articulate and suss out a methodology for myself. This would also help me analyze my own pieces and the work of other composers more productively, I believe.

Thanks in advance for your ideas!

You need to be a member of Composers' Forum to add comments!

Join Composers' Forum

Email me when people reply –


  • I actually just thought of something I'd like to add to this. One of my major issues is with relying on playback through Sibelius to evaluate how things are working structurally rather than my internal ear and also, I guess, a better knowledge of structural principles. I've actually read that Michael Daugherty has relied heavily on MIDI playback to work things out, even in the early days in the eighties when MIDI renderings were even worse than they are now. He's actually a great orchestrator and a pretty good composer overall. However, I find the repeated listening to MIDI playback to be, not only annoying, but also very misleading. Of course, there is a corrective once a piece is heard played by live musicians, but it's very hard to imagine things being played differently (i.e. in terms of dynamics, phrasing, etc.) when you are constantly listening to a MIDI rendering.

    Have any of you struggled with this and what have you found to be some solutions? In an ideal world, I wish I could compose without any instrument at all.

  • I have always found MIDI renderings to be very distracting when I'm in the initial stages of composition. It gives you a false sense of security that you have achieved instrumental balance, when in reality you have only achieved coincidental "balance" among the generally-poor quality MIDI sound patches (that unfortunately, more often than not aren't consistent with each other, so for orchestral pieces a "properly" instrumented passage would sound off on MIDI, and what sounds good on MIDI may turn out to be horribly off-balance in real-life). The mechanical quality of most MIDI renditions can also throw you off by forcing you, subconsciously or not, avoiding melodies that a human player would be able to play in a beautiful way, but that is completely butchered by the MIDI. So you end up not writing that beautiful melody, but something else that may sound nice on MIDI, but only mediocre in real-life!

    Generally, I tend to compose in my head instead, occasionally using the piano to help me hear the pitches more accurately (or to catch any wrongly-notated notes). I may use the MIDI rendition occasionally to give me a rough idea of what something might sound like, but especially recently, I found myself avoiding MIDI playback where possible, in order to avoid being unduly biased by it. Even for piano pieces, I'd compose at the piano and work things out there, and enter notation on my computer without any audio feedback; I'd only listen to the MIDI every now and then to make sure I didn't wrongly notate something, but I wouldn't listen to it too much unless I've already finished the piece.

    Once I finish the overall piece, sometimes I do try to make the MIDI sound a little more tolerable, but that usually only works well for certain kinds of pieces. For others, it would take so much work just to make it sound tolerable that it's simply not worth the effort.

  • Hi Michael,

    You and I share a common goal. I also, as you do, "like to keep the material unpredictable." I mostly compose in an organic style, by which I mean I let the flow of the composition determine its own direction as it goes along and do not concern myself with following ABA structures or the classical structures such as fugue or sonata. In short works, a single idea I find sufficient to structure a work - it can be a slow or a fast piece, and not have any variations on that central idea and still be successful. In longer works I seek contrast as the way to keep the piece interesting. You can't have IMO a single idea that goes on for many minutes before becoming boring. In a longer work, if the piece starts out slow, I want to contrast that with faster sections. If a piece starts out with traditional harmony, I may want to vary that harmony into something more outside of the beaten path later in the work. One other point, theoretical knowledge is good, but not every good composer has needed it. Erroll Garner, the wonderful jazz composer (mostly known for his 60s hit "Misty") couldn't read music and had no formal training. Here's one short work posted by me on the forum which I think illustrates a single idea and which is also organic in the sense that it is both unpredictable and at the same time coherent in the sense that it has a recognizable melody (in the bass) which returns, develops, and has its own structure which leads to a (I hope) satisfying conclusion -

    Underground River Ride

    Underground River Ride
    This is a piece for solo piano, inspired by the Indiana Jones movies and other adventurous shows like Jonny Quest. Imagine a chase scene with bad guy…
  • As for your original question about form, I wrote a long reply (perhaps far too long :-P) but pressed the wrong button and the computer swallowed it. :-(  But anyways, the gist of it is that at one point in time I decided that I wanted to write something in sonata form, but found that I couldn't, because it's not just a matter of following the letter of the form, but the themes themselves had to be be suitable -- they have to "want" to be in sonata form, otherwise the result won't "work". From that, I learned that the music itself "wants" to be in a certain form. Or conversely, if you want your piece to be in a certain form, you have to shape your themes in a way that would "naturally" result in that form.

    Using sonata form as an example, the first subject must have a certain kind of inherent "instability" to it; it has to "want" to end up in the dominant key (or some other key, depending on how modern you want to be). A melody that naturally cadences in its home key would be unsuitable for this purpose, because any attempt to force it to modulate to the textbook dominant key (or some other key) would sound forced and unnatural -- the music wouldn't "flow" correctly. The second subject, in contrast, tends to have a stronger sense of "stability" -- it "likes" to cadence in the key it started in. However, because the instability of the first subject has led to the second subject starting in something other than the home key, when the second subject ends the audience subconsciously feels like there's something still missing -- there's still an unanswered question in the music. As a result, they naturally expect the music to continue. If the music were to end there, they'd be left wondering how come there wasn't more. Conversely, if the first subject had a strong conclusion, then when the second subject begins the audience would feel like it's something "tacked on" to the end, and they would feel like the music is dragging on for no good reason. Or if the second subject cadences in the home key, then the development and recapitulation would sound redundant -- that cadence has already resolved all of the musical questions, so tacking more stuff on feels like it's an appendix rather than part of the story.

    Outside the realm of classical music and sonata forms, the same principles apply. If your themes "want" to go a certain direction, eventually you have to let them go where they need to go, and that will result in a certain form. Conversely, if you want your piece to be in a certain form, then the themes should be shaped accordingly. You can't just throw in random motifs and then after the fact try to shoehorn them into a particular form -- the result rarely sounds good. Somebody once said, that the best music is the one where the entire piece has a sense of inevitability -- the unfolding of the entire piece is "inevitable" because the initial themes "wanted" to go in that direction.

    This "want" or "drive" in the themes can come in many different forms -- a tendency towards a different key, or an unresolved discord that makes the listener want to hear a resolution (that doesn't quite seem to want to come until the very end), or various other kinds of tensions. An opening motif itself could have an inherent tension that seeks resolution -- you'd then basically use the rest of the piece to "answer" that tension.

  • Thanks for the observations, guys. I don't have it in me to give more than a brief reply right now, but I'll say these ideas are very helpful. I think the challenge as a modern composer is that there are so many more vectors that determine large-scale musical coherence, apart from relationships between closely-related keys, as was the case in the past. (Pardon the wholesale exclusion of non-Western music.)

    But now we have a century-plus to look back on, where composers organized music in completely different ways that have yet to be "universally" codified, such using timbre, density, modal modulation, etc. to serve as the overarching framework. H.S., I like that you brought it back to the more abstract notion of tension vs. release. In those terms, we can still draw upon the older forms when we write.

    I guess what is in the back of my mind is, some day I would like to tackle longer form and larger scale works, such as a multi-movement orchestra piece. Like an architect, I want to make sure I have laid an adequate foundation in terms of form and that I'm not using materials that will collapse due to a basic ignorance of their elemental qualities. Perhaps, I need to first write more smaller scale works to build my confidence in that respect.

  • I too, have not so far not yet "conquered" multi-movement pieces. I did write a quartet in 4 movements, but unfortunately it's unfinished because I didn't know how to end it! It's also very elementary; each movement is relatively short and does not have large-scale structure. The movements all link to each other (they end on dominant V chords rather than a perfect cadence), so arguably it's a single movement in multiple sections rather than multiple movements proper! I'm also unhappy with its overall quality, except perhaps for the bouncy, light-hearted 2nd movement. Perhaps it's like you said, the material collapsed because I wasn't experienced enough to shape them the way they needed to be in order to "work" for a 4-movement piece.

  • I have an online composition tutor at the moment, and he is very insistent on structure. My last completed piece was also a result of this insistence, and I found that,despite my initial resistance, it helped the process enormously.

    In this earlier piece (a 9-movement work), all the themes were stated in the first movement, and the remaining eight were variations, excerpts or permutations of these 1st movement themes. It really did help to coalesce the whole work, and gave it a real sense of unity.

    In the current work (5 movements) there is a theme for mvt 1. The second movement theme uses the first theme backwards in one part, and the same theme backwards in the second part, but only using every second note. Works well.

    The third movement (the 'central' movement) uses a completely new theme (theme 2). The 4th movement will be variations on theme 2, and the final movement will be variations on a combination of themes 1 and 2.

    This approach, much to my surprise, does nothing to impede the organic flow of creativity. Rather, it gives it more focus along the way, and gives you a sense of the overall 'aim' of the piece, which actually seems to help the process.

  • The best statement I ever heard about the art of movie-making, which I think also applies to long-form music, is from William Goldman, screenplay writer for The Princess Bride, who said (paraphrasing) - "the ending of a good movie must be both surprising and inevitable" - in other words, you must not be able to guess the ending, but when it happens, you should say to yourself "of course, how could it be anything else?" This sentiment is always in the back of my mind when composing longer works. What better result could a composer seek than to get such a reaction from the listener?

    H. S. Teoh said:

    Somebody once said, that the best music is the one where the entire piece has a sense of inevitability -- the unfolding of the entire piece is "inevitable" because the initial themes "wanted" to go in that direction.
    Approaches to structure
    Hi everyone, I often take forever to complete pieces, so my mission lately has been to find ways of being more methodical with my writing. I don't me…
  • One of my favorite examples of this is when a piece ends with the sudden reappearance of its opening motif. Off the top of my head: Dvorak's 5th symphony (last movement draws to a close, then at the climax the opening motif of the entire symphony suddenly reappears to round it off), Shostakovich's 7th symphony (climax of last movement is the victorious reappearance of the melody that opened the entire symphony), etc..

    (Of course, now you'd go "yeah that's so predictable, make the opening motif reappear at the very end"... but there's a big difference between just recycling this old trick, and doing it effectively. If you do it right, the listener both doesn't expect that to happen, and also feels it "fits", and isn't just another hackneyed rehash of an old technique.)

  • Hello Michael (Michael Lee Moore),


    You observed,


    “I find the repeated listening to MIDI playback to be, not only annoying, but also very misleading. Of course, there is a corrective once a piece is heard played by live musicians, but it's very hard to imagine things being played differently (i.e. in terms of dynamics, phrasing, etc.) when you are constantly listening to a MIDI rendering. … Have any of you struggled with this and what have you found to be some solutions?”


    MIDI’s played by ordinary playback software (like QuickTime) do usually sound quite awful, in my opinion. However, you say you are using Sibelius.  Do you play back most of the music you create via computer, with a straight MIDI using QuickTime, or do you play it back, most of the time via some kind of composition software, like Sibelius or Logic Pro


    Repeated listening, as you are composing, enhancing and correcting is (I think) a very rewarding activity, as long as the listening is not rote.  I am simply talking about my own experience, but I think there may be “solutions,” or at least different approaches, that might be helpful.  I use Logic Pro X now, but I found that even the earlier versions of the program had many features that facilitate composing in impressive ways. 



    I’ll post some remarks and questions about things people might be able to do, and maybe people will be able to comment on them. Some of the things I say below may seem very basic, and perhaps they are mostly things you do, or have done, so if that’s the case, I apologize.


    (1) Can composers try to play the sound file a different way each time they listen? I attempt to play it differently each time, by adjusting parameters frequently.  Whatever one’s initial concept of the composition, are there not things to do which help one see the piece in a new light (or in dozens of new ways)?  In the early stages, perhaps it is wise, and useful, to adjust the tuning several times, and play the work in several ways, including equal temperament, fixed tunings, 12-tone Pythagorean (even in an Arabic tuning or Siamese tuning), and in dozens of other tunings. Isn’t this beneficial, just so one can hear different aspects of the melodies and harmonies that one might ordinarily miss?  Ultimately, one will set the tuning to a standard mode, of course.  [Logic X actually has a new and interesting feature that lets one change any given sequence of notes-- both melodies and harmonies included-- from whatever key it is in, to a major, minor, chromatic, Lydian, Mixolydian, Dorian, Locrian, etc. etc. sequence or mode …] Won’t experimentation with the harmonies and intervals keep the piece fresh? Of course, one always holds on to the originals and backups so nothing is lost.  An insight may be gained, so one can go back to the original and adjust accordingly.


    (2) We begin, usually, with an idea about orchestration or what instruments will be used.  [I like what you said, “In an ideal world, I wish I could compose without any instrument at all.” – and in a way, this can be done: we can compose with ALL INSTRUMENTS, which is a bit like not composing with any]. I find it is useful to listen and re-listen to the same passages with a multiplicity of different instruments or instrument combinations.  Even if the original melody is written for violin or a string ensemble, isn’t it sometimes useful to listen to it played by a host of other instruments?  One can try harp, piano, brass instruments, various woodwinds and even percussion, just to see if there is some aspect of the thematic or harmonic structure that needs changing, or to observe whether there is some possibility that might have been neglected.  Logic has something called “sound sculpture” that allows you to actually create instruments, or alter the sound of normal instruments; and the EXS settings of each instrument also allow you to alter many parameters, in ways that make the composition process extremely interesting.   I don’t know anything about Sibelius composition software, but doesn’t it have similar features?


    (3) Sometimes a melody appears or sounds rote, or a harmony becomes rote, after repeated listening.  Isn’t it useful to experiment with inversions, retrograde sequences and retrograde inversions?  One can use the software to create sound possibilities not easily imagined in advance, or heard in one’s head. Can’t we subtly adjust many aspects of the note sequence, either speeding it up (using subtle tempo shifts, or doubling, or halving the tempo)?  Can we make the piece sound less mechanical?  I think the software gives us any number of methods to do this, and many other things. 


    Do we ever want to free ourselves from “the tyranny of the bar?” This can be done by direct personal intervention, handling each note carefully using manual means. There is something called “humanization,” which slightly alters the duration, location, and velocity of any note or note sequence. Is randomization ever an option? This can be applied to as large or as small a portion of the music as one likes.


    (4) If we think of the MIDI as something fixed (as it is, according to the old-fashioned “event list”), we can get very frustrated with it.  But I find that even small adjustments with the MIDI within advanced software can produce very dramatic and often positive changes, and supply us with new insights.  Can’t just a slight change in the dynamics of one or two notes create a huge difference in the mood and meaning of a melodic sequence?


    (5) The main complaint about the computer-generated file is that it often does not sound “natural” or that it sounds too mechanical.  Won’t this become less and less of a valid objection as time goes on? I don’t see why not. Won’t the problem be resolved as people become better at using the software, and as the technological changes increase the verisimilitude of the product?  I think that seems likely. 


    Do other people have ideas on this aspect of the topic proposed by Michael?  I am suggesting what are only some of the most obvious means of making the review of a MIDI or a sound file interesting and creative, from the point of view of composing.  I am discovering many more as time goes on.  Can people say more, having to do with the automation functions, involving the “read,” “touch,” “latch” and “write” buttons, which can be found in Logic (though I don’t know what their equivalents are in Sibelius)?


    Isn’t there sufficient leeway and flexibility in these systems to make composition an activity that is both satisfying and creative?



    Michael Lee Moore's Discussions
    Michael Lee Moore's Discussions | Meet composers from all over the world. Share ideas, thoughts, tech insights, job opportunities, and business knowl…
This reply was deleted.