Posts Tagged ‘ Structure ’

Nested Structures

Nested structures are quite simple to understand, but can add new levels of order and structure to your music.

If we start with two basic structures:
Binary: A B
Ternary: A B A

Nested structures refers to the idea that each of the structure “elements” (A, B, whatever) can actually be (or have) a whole structure in itself. This can be the basis for develping more complex structures from simple ones. For example:

We could choose ternary for our overall structure (ABA). But if we split it up further – replace A with the binary structure ab, and replace B with the ternary structure cdc, then we end up with the overall structure ab.cdc.ab. Read through this a few times if you didn’t quite get it.

Now think about taking that another layer deeper. You can keep nesting structures until you get down to individual phrases, gestures, motifs, even notes!

Also, consider that there are many more possibilities for “basic” (or primative) structures. As well as binary and ternary, there’s also rondo (ABACADA – commonly chorus,verse,chorus,verse,chorus,etc), sonata (A B <development> A B’)


Usually, there’s two ways to approach this: top-down and bottom=up. A top-down approach would be very similar to my example above – start with an overall structure, and then split it up into smaller and smaller pieces, stopping when you feel that you can easily populate a single piece. This is a “divide and conquer” approach.

A bottom-up approach would be the exact opposite – start with several very small pieces, then arrange them into larger and larger structures. This is very easy to do in a sequencer, where you can develop a few one or two bar sections, then copy and paste them in various orders and configurations.

Personally, I usually use a combination of the two. I build the piece bottom-up, but when I’m doing it I have a mental “plan” of how I want the entire piece to turn out.
By approaching composition in this way, you can create a piece with a very high level of coherence and order. Each section will fit exactly in its place, and repeated sections can give a certain unity – without having to resort to a simple verse-chorus-verse-chorus (or similar) structure. It’s also an easy way to add complexity by using relatively simple ideas.

Of course, the fun begins when you combine nested structures with techniques for subverting nested structures. Build your piece as usual, but add interest with variations, interruptions, twists and turns, bizzare trips to strange places. These kind of subversions are usually much more effective when you start with something with very high coherence and order.



Transition sections that are too long…

If you have a transitional state between two sections that have similar rhythm, pace, tonality, register, texture, etc; it doesn’t take much time to move between them. However, if you try to move between two very different sections, more time will be required (for the same rate of change). The time required to move from one state to the other is somewhat proportional to the “distance” (or difference) between the two states.

Sometimes it’s important to have a long transition in order to put some distance between two sections.

The problem is that a linear transition gets boring very quickly. Alinear transition is one that just goes straight from A to B. For example, if you have a dark and resonant synthbass sound in section A and a bright and harsh synthbass sound in section B, then a linear transition would be one that (among other things) just gradually opens the filter, like a straight line – slow and predictable.

And that’s exactly why a linear transition is boring – because it’s predictable.

If reducing the length of the transition is not an acceptable solution, another alternative is to break it up.

  • Change the “curve” of the transition – Rather than move predictably from A to B, perhaps start the transition with a very low rate of change, and gradually increase (the rate of change)… This would have an effect of the transition section intially not sounding like a transition – but more static (unchanging). Depending on the context, this can either create tension (possibly good) or boredom (probably not good). Slowly we hear more an more changes happening (the changes speed up), “climaxing” with a very (or more) dramatic change at the end – just as the next section begins.
  • Put breakpoints in the transition – Rather than simply starting at “State A” and moving towards “State B”, you could add animation between the two points. For example, the transition moves from A, but when it gets about halfway between A and B, it goes back to A again. Then we start moving towards B again, but this time get about 75% before moving back again. and maybe a third time we finally reach B. This can be particularly interesting if the listener already has some knowledge of where the transition is going (s/he has already experience B in some form). another interesting effect of this is that it can distort the listener’s sense of time as well, but this is more difficult to control and excecute well. This technique could also be extended by including “fake” transitions to a different state (not A or B) – this will also manipulate the listener’s expectations.
  • Insert contrasting sections in the transition – Four bars of transition, then four bars of something completely different, then the four next bars of the transition, then another four bars of something completely different, etc etc etc. Of course you can vary the length of each “mini-section”, you could have each contrasting section be something relevent from another part of the piece. You could effectively interleave two different whole sections – even have two transitions coming from different states but both moving to the same final state.

Of course, these ideas aren’t restricted to transition sections – you can use variations of these ideas to add interest to any long section. You could even apply them to only some elements of the arrangement (for example, make the synth parts go through a transition, but keep the drums and bass consistent).


Beginning, middle, end

Normally when we compose a piece of music, we are working on it in a non-linear fashion. That means we can work a little on the start, then work on the end, then maybe add a new section in the middle, whatever. Also, our perception of the piece is non-linear – being so intimately involved with the piece (and its construction), we usually know the entire piece by memory. That gives us (the composers) the ability to compose parts of a piece in the context of the rest of the piece.

Your listener, however, will have a very different experience of the music. As an artform, music is particularly interesting because it exists in time. You listener will listen to your piece by starting at the start, listening through each moment once, and stopping at the end.

Consider that the listener will also have no (or at least, very little) knowledge of the piece before listening. S/he will begin the listening experience knowing nothing, and gradually (and linearly) learn more about they piece as it is experienced. I like to think of this as an “unravelling” or “unfolding” of music – as the listener experiences the piece, it is being revealed, opened up.

This observation has interesting implications for different sections of a piece.

The beginning is significant because it “introduces” the language of the music to the listener. When you listen to a piece of music, the beginning is the first thing you hear – and thus, it is what influences the expectations that you have for the rest of the piece. It is what sets the context for the remainder of the listening session. When composing the beginning of a piece, consider that this is the first thing your listener will hear.

The ending is (in this respect) the complete opposite – the listener hears it in the context of the entire piece. By the time the listener gets near the end of a piece, s/he has travelled through the “journey” of the music, and (hopefully) understands the language[1] of the music. When composing the ending, consider that the listener hears this after hearing the entire piece through once.

The middle of a piece is also interesting, because this is (usually) where the “scene has been set” – the listener has some idea about what the laguage of the music is, and what to expect for the rest of the piece. Most well-written pieces use a/the middle section to develop and enrich the listeners understanding and experience of the world you (as the composer) have created.

All this, of course, doesn’t mean that this is the way it has to be, or that this is the (only) way to compose “good music”. As the composer, you are free to subvert the rules or discard them completely. However, understanding how a listener listens to a piece will (hopefully) help you make better informed decisions during the composition process.


Sections of variable length

Often I’ve found that using sections of “metric” lengths (four bars, eight bars, sixteen bars) can often give a piece a very rigid, predictable pace. No matter how exciting or interesting the actual musical material is, sections of metric length can really weigh a piece down.

This is because the listener knows (or can guess fairly accurately) when each change will occur. In her/his mind, the listener has heard a sigificant amount of the piece before it’s actually been played.

Let me give you an example. Let’s say that so far, every section has been sixteen bars long, and it’s very obvious whether each section is static or transitional. Within a few bars of hearing a particular section, the listener already knows what the rest of the section sounds like – sometimes to the point of not actually having to hear the remainder of the section. This is the point at which the listener becomes distracted, starting to talk, or getting bored.

A particularly effective way to reduce this effect is to use variable section lengths. Instead of making each section a “metric” length (four bars, eight bars, sixteen bars, etc), the idea is to make them “odd” lengths. This has two implications:

  • The listener will not be quite so sure how long each section will be. In fact, (if done well) sections will often end/change earlier than expected or later than expected. This can be taken advantage of to highten expectation and excitement.
  • The internal structure of each section will be more “fluid”: In sections of “metric” length, we tend to break them up into smaller bits of even length. For example, if we have a section of sixteen bars, we might very easily put in eight chord changes, one every two bars; or four chord changes, one every four bars. If we have a section with an “odd” length, it forces us to be more creative with the internal structure. For example, if we have a section that is thirteen bars long, we might split it into three groups of four bars, plus one; or four groups of three bars, plus one; or three groups of three bars, plus four bars… or anything else.

How you come up with the lengths is up to you. I composed a piece a several years ago where each section length was a Fibonacci number – the sections were all lengths like 5, 13, 21, 34, etc.

Another piece I composed had section lengths chosen by rolling dice.

Of course, it doesn’t have to be random. You might choose prime numbers, or the date of every Monday in the year, or anything else. You could even choose the lengths as you compose the piece, depending on the flux in the piece.

It’s really just about making the sections have lengths which aren’t even multiples of four or eight.



Buildups often require particular attention when composing. A buildup section is one immediately leading up to a point of high energy.  Commonly this is the climax of the song – the most important part of the song. The buildup is critical because it has to lead up to the climax in a way that maximises its effect. This is best done by enhancing the listener’s sense of anticipation and expectation.

I’ve written about expectation in this post, but that alone is not enough. For a stronger effect, also consider sequences, precedent,  linear movement and transition.

Sequences are repeated patterns in music. As explained in the post on expectation, sequences play an important role in setting the listener’s expectations. Worth considering, though, is using sequences at different levels. For example, you could use repeated patterns within the buildup to create a cyclic effect.

Precedent applies this idea to a wider scope – whole sections. You could also make the buildup itself part of a larger sequence. By having a smaller buildup leading up to a smaller climax earlier in the song, you heighten the listener’s expectation of a bigger climax at the end of a bigger buildup.

Linear movement is important in a buildup. Many song sections are static – they stay the same throughout the section (the same level of energy, the same types of sounds, the same density, etc). The buildup, however, works best if it is in motion. Typically, this works best if the buildup is gradually getting louder, more energetic and more complex. The buildup section might start quite subdued and understated, but at the end it might have enough energy to meet the climactic next section.

Transition is related to linear movement. It might help to think of the buildup as a transition section between the climax and whatever precedes the climax. Using the buildup as a transition section also helps glue the song together because it links the two adjacent sections (rather than sounding as a separate section on its own).

Another useful technique is to introduce a short “gap” in between the end of the buildup and the beginning of the climax. As explained in the last section of the post on expectation, this pushes back the climax and increases the sense of expectation and anticipation in the listener.


What makes structure work?

After experimenting with different approaches to structure, you will begin to vary standard structures and start to think about developing your own approaches to structure. You might start to wonder – what makes structure work? What separates an effective and satisfying structure from an ineffective one?

Contour and proportion

Contour and proportion are about the overall shape of the structure. A structure with good contour is one where the overall rise and fall of tension and excitement makes sense. This means it can be understood by the listener as having a shape that can be followed. The obvious and most common shape is one where the song begins with low excitement, gradually increases to maximum excitement about 2/3 through, and then ends at minimum excitement again. This isn’t the only shape that works though! Another shape that makes sense is one where the excitement is greatest at the start and the end, but the middle section is quiet and subdued. Several Juno Reactor songs from their album Labyrinth have a contour like this.

Proportion goes hand-in-hand with contour. At a slightly smaller scale, proportion is about the lengths of each section. A structure with good contour is one where each section is just the right length – not too long or too short. Poor proportion is usually caused by sections that are too long without enough change to keep them interesting. There’s no easy rule to help you determine the right length – you have to use your experience and judgement. Shorter sections can be useful for increasing excitement and expectation because they make it feel like the song is moving along at a quicker pace. Longer sections are useful for building tension because the listener is expecting a change that is postponed, or for maintaining and emphasising a hightenened level of excitement during a climax.

Expectation and fulfillment

Expectation and fulfillment go hand-in-hand when approaching structure. Expectation, as you might guess, is what happens when a listener thinks s/he knows what is going to happen next in the music. This expectation is shaped by many factors you can’t control, such as personal taste in music and genre norms. A factor you CAN control though, is repeated sequences. For example, if you have three sections – A, B, C – and arrange them in your song as A-B-C-A-B-C-A, the listener will excpect section B to follow. As a more real-world example, standard song form begins with verse1-chorus-verse2-chorus, after which the listener naturally expects a third verse to follow.

Fulfillment is what happens when the listener’s expectations are met. In light of the above explanation of expectation, the listener experiences fulfillment when the section that logically follows is the section that actually follows. Note – this is not always a good thing. If the listener’s expectations are fulfilled too much the song is percieved to be predictable and boring.
Instead, building expectation but not fulfilling it helps add surprise and interest. It can also support a sense of development and movement in the music. For example, standard song form begins with verse1-chorus-verse2-chorus, at which point the listener naturally expects a verse3. Instead, there is a bridge – new material that surprises and adds interest, and also gives the song a sense of development by increasing the musical scope (adding musical material to the song).


Coherence is about the amount of musical material in a song. An easy way of thinking about this is to consider the number of different sections (or melodies, or themes, etc) in a song, as well as the overall length of the song. A song with a high level of coherence will not have much musical material – it might have fewer different types of sections, or its sections might be very similar. Conversely, a song with a lower level of coherence will have a lot of musical material – either more different sections or more variations. Some level of coherence is necessary in music – in order to give the song a distinct musical identity and so that each part sounds like it belongs to the same whole. Too much coherence, however, will make a song boring and repetitive.

Time is also a significant factor contributing to coherence too. Given a certain amount of musical material (say, for example, three different sections), you can increase coherence by increasing the overall length of the song. Similarly, you can decrease coherence by shortening the length of the song. This is an often-overlooked approach. If you’re working on a song and you feel like it’s too boring and repetitive, try shortening it instead of simply adding new material. Likewise, if you have a lot of musical material (many different sections or musical ideas to organise) and the song is feeling like it doesn’t have a distinct musical identity, try making the song longer. This will let the music breathe a bit more – allow the musical ideas to expand and develop.

As with all composition techniques, using them in extreme is usually not the best approach – some judgement is required. And as with all composition techniques, practice is necessary for mastery! You won’t get the hang of this first time around – give yourself a few songs to experiment, to develop your own sense for how it all works.


Transitions between sections

All this talk of structure revolves around sections – necessarily so, because we’re talking about organising a large block of time, and the most common way of doing this is by subdividing into smaller sections.

No matter how you organise your sections, you will still have a skeleton of a song that consists of several sections of various lengths, one after the other. Without any transitions between sections, each section will simply stop as the next begins. The effect will be similar to that of changing channels on a television – abrupt and unsophisticated.

To make a transition between sections work, you must make something of it. Articulate it in the music, make a point of the change. Necessarily, there are one of two approaches you can take with each transition – a smooth transition or a contrasting transition.

Smooth transition

A smooth transition is one where the first section smoothly moves the listener into the second section. A common example of this is where the second section is fuller and more exciting than the first section, so the end of the first section has a build up into the start of the second section. Similarly, if the second section is slower or sparser than the first section, the end of the first section might pull back or slow down before entering the second section. The second section might even continue to get sparser in the first few bars.

At an extreme, the transition between two sections might be long enough to be treated as its own section. That is, a whole section in the song is dedicated to transitioning from the previous section to the next.

Contrasting transition

By contrast, a contrasting transition is one where the change from the first section to the second is marked and noticable. It doesn’t have to be sudden, but it does rely on the two sections being quite different. An example of this might be the sudden jump from a sparse and soft introduction to a song into the full and busy main part of the song. Another example could be a jump from the second chorus of  a song into a contrasting bridge section.

At an extreme, a deceptive transition can be used to further emphasise the contrast. An example of this could be where the first section ends by building up as if the second section is louder and fuller, but instead the second section is suddenly quiet and soft. Another example could be where the first section ends by slowing down and pulling back (perhaps even pausing) before the second section suddenly bursts in.