Why Your Chorus Doesn't Work

one of the hardest things about songwriting is transitions. you can write a great verse and a killer chorus, but when you try to connect them, it just doesn't… work. something feels off. you're looking for a big, exciting release, and it feels like it should be there, but somehow, it's not. the song becomes less than the sum of its parts. a well-executed section transition is one of the clearest signs of an experienced arranger, and while there are lots of things to consider, there's one you're probably not paying nearly enough attention to: texture. let's fix that. (tick, tick, tick, tock) texture is the music theory name for the overall sound palette of a song, and one of the reasons it's so hard to talk about is that there's so many options. compare that to, say, harmony. in popular music, pretty much everything uses the same 12 notes, arranged in only a handful of different chord qualities, so most progressions are built from the same, I dunno, 50 chords or so. certainly less than a hundred. and even that's overselling the difficulty, since many possible chord motions are never used, and the same progression in different keys sounds basically the same. this narrow possibility space makes it easy to spot patterns, build models, and develop a shared vocabulary for what different chords do. but texture isn't like that. artists don't stick to a predefined list of sounds. or, I mean, sometimes they do: back in the '80s, digital synthesis was a brand-new technology. the earliest synths were really hard to program, so across genres, tons of hit songs were built off just a couple presets on the Yamaha DX-7. most notably, their recreation of the Fender Rhodes, called E PIANO 1, used by artists ranging from Earl Thomas Conley: (bang) to Rick Astley. (bang) and then there's the Roland 808, whose synthetic drums defined a generation of hip-hop. (bang) but these stock sounds are the exception, not the rule: modern Digital Audio Workstations make it trivial to create new synth tones, and historically, the mix of amp settings, effect pedals, and just physical differences between instruments creates near-infinite possibilities. and even if none of that were true, we're still not even talking about texture. this is all timbre, the sound of an individual instrument. texture is the acoustic conversation between all the instruments, adding yet another layer of complexity to the problem. this is why music theorists tend to think about texture in terms of layers, groups of instruments that work together to perform a certain job and occupy a certain portion of the mix. according to Dr. Allan Moore, there are four main textural layers, which I'll call the beat, the bass, the melody, and the filler. I already made a video about these, which I'll link in the description if you want a full explanation, but to sum it up, the beat belongs to unpitched percussion, like a drum kit, and its job is to explicitly define the metric structure of the piece. the bass is for low instruments that bridge the gap between the rhythm section and the harmony. the melody is higher, featured instruments, including vocals, but also guitar solos, synth riffs, and so on. basically, if you can easily hum it back from memory, it's probably the melodic layer. and finally, filler is all the stuff in between. it provides the sonic backdrop that frames the rest of the music. if we want to understand section transitions, these textural layers are hugely important. adding a layer, removing a layer, or even changing what's happening inside one, are all really clear signals that the music has moved somewhere new. and the importance of textural changes has even been growing in recent years: as the influence of electronic dance music spread through the pop industry in the '90s and 2000s, it brought with it a new, heavy emphasis on texture as its primary driving force, and by now that influence has worked its way into pretty much every mainstream genre. as an aside, this is part of why I'm not very convinced by arguments that modern music is getting less complex: any decrease we might have seen in harmonic complexity is, to me, more than offset by the wild texture plays happening in basically every pop song. it's all there. you just have to know what to listen for. anyway, back to the point. if we want to write good section transitions, we first have to understand the sections themselves. what are their textures trying to accomplish? to borrow a term from Dr. Asaf Peres, what is their sonic function? most pop songs can be broken up into three main chapters, each ending with a chorus. chapters 1 and 2 start with a verse, and maybe a prechorus, while chapter 3 starts instead with a contrasting bridge. this isn't, like, a rule of music or anything, lots of songs have other forms, but I'm just gonna stick to this one for now 'cause it makes things easy to see. besides, many of those forms are just variations on this one, so a lot of the ideas will still apply. point is, each of these chapters has a different job. Dr. Peres calls them the set-up, the build-up, and the climax, and from those names, you can probably get a sense of what they do. the set-up establishes the basic energy level, the build-up starts to increase it, and the climax delivers the knockout. but it's not quite that simple. within each chapter, we can find these same three functions again on a smaller scale. the beginning of each chapter is like its own mini set-up, that then builds up to the local climax of that chapter's chorus. this gives us an overall dynamic contour that looks something like this

with peaks and valleys to pull you in and keep you interested. exciting moments are spread throughout the song, but they're also all building toward that one final release. knowing where you are in that curve is essential if you want to make your section transitions work. there's a lot of ways to do that, but for this video, I'm gonna focus on what Dr. Emily Schwitzgebel calls textural cues, little texture changes that happen right before a transition. these work because your brain loves making assumptions. even when you're not paying attention. especially it's constantly processing all kinds of information and making guesses about what's gonna happen next. when those guesses are right, it feels good. textural cues are a subtle way of intensifying that reward by helping your listener's brain to assume the right things. they're like a little signpost that says "hey, the song's about to do this", and that signpost works best when they're right next to the actual change. that's thanks to a phenomenon psychologists like Dr. Peter White call the extended present. basically, the amount of time that still feels like now. this actually comes up a lot in my work: when I draw these doodles, I try to keep my hand in for no more than a couple seconds at a time, because if I don't, if I just keep talking while slowly drawing one single thing that you can't really see yet because my hand is in the way, it starts to feel like you're waiting for the finished picture, rather than just watching it appear. there's no clear definition for exactly how long this extended present is: according to White, different scholars have given estimates ranging from 2 seconds to around 10, although he argues that it's actually less about the length of time and more the amount of information being processed. if less stuff is happening, the current moment lasts longer. but no matter how exactly we define it, research by, among others, Drs. Roni Granot and Nori Jacoby has shown that cues work best when they exist on the sorts of local timescales that might fit into the extended present. and that makes sense: the closer they are, the easier it is to connect them, and the more confident you can be that your correct prediction wasn't a fluke. if you place a cue for the chorus at the start of the verse, then by the time the transition actually happens, the signpost will have been forgotten. so what do these cues look like? well, Schwitzgebel's paper describes four common approaches. these come from her study of post-millennial pop, because, again, that's probably the most interesting place to look for texture stuff. but she does keep them fairly broad. some version of each of these can be found in a wide range of styles. the first one is Rhythmic Acceleration. this isn't, like, a tempo thing, where the actual song speeds up. that's not really texture. instead, a rhythmic acceleration cue is a temporary increase in the rate of attacks in the beat layer. that could mean a higher pulse rate for programmed drums: (bang) or it could just be a drum fill. (bang) and here, I want to address an important point: these are textural cues, but they're never only texture. this, for instance, could easily be understood as the build-up and release of rhythmic tension without me ever having to say the word "soundscape" again. and that's kinda how texture works: it's the glue that binds all the other musical elements together, so while it's clearly important, it's practically impossible to isolate. everything is texture, but nothing is just texture. but I still think this counts, for two reasons. first, fills often involve drums that weren't otherwise being used, like toms or cymbals, while sequenced accelerations can pitch-shift the samples as they build. that introduces new timbres to the beat layer and new timbres means different texture. that's the easy one, but more important is the second reason, which has to do with why we separate the texture into layers in the first place. because again, a layer isn't just about the sound, it's also about the job that sound does. the beat layer's job is to explicitly define the metric structure. that's what we expect from the instruments we put there, and this rhythmic acceleration is a temporary violation of that function. the beat moves out of sync with the music, implying a faster pulse than is actually happening. when it snaps back into alignment on the transition, you don't just get a rhythmic release, you get a texture one too. that said, our next cue is probably the closest we're gonna get to something purely textural: Layer Withholding. this one's simple: you just cut out one of the layers. that's most often the beat layer: (bang) but it could also be the filler: (bang) or, really, anything else. like before, this works by momentarily subverting the layer's function, leaving you without whatever structure it was supposed to be providing. what exactly that is depends on which layer you cut, but no matter what, suddenly withholding a sound that you know is supposed to be there is very obvious signal that you're about to hear something big. and this one's extra useful 'cause you don't even have to actually change textures: taking something away, then bringing it back, can still create a really effective transition, without needing to write anything new. the third type of cue is the Melodic Soar, which doesn't actually take place in the melodic layer. instead, this is a rising melodic gesture in the filler. depending on style

this might be a filter sweep: (bang) a guitar slide: (bang) or something else entirely: the sound of the filler layer is really flexible, so it's hard to list all the possible instruments you could do some sort of rising line with. no matter what, though, the effect is the same, sweeping you up and building excitement that pays off in the transition. it's like someone grabbing your arm and saying "here we go! " right before the rollercoaster goes over the hill. and last, we have Upbeat Activities, which is any quick rhythmic event that adds a new timbre and happens on the last beat right before the transition. this can happen on any layer, but it's most common in the beat. (beat) these serve as an unexpected jolt that opens your ears so you're really listening when the transition hits. all these cues have to be short, but in this case the shortness is the whole point, compressing the set-up and payoff into one single musical event. you're bouncing off the upbeat and ricocheting into the new soundscape. those are all the cues from Schwitzgebel's paper, but she said it wasn't meant to be an exhaustive list, so I'd like to add a couple of my favorites as well, starting with what I'm calling the Preempt. this is where one of the layers jumps to the timbre of the new section early, most often the vocals: (bang) but it can happen anywhere, and I'd differentiate from the upbeat activity cue in that it's usually at least a bar or two long. it's kinda like applying a blur to the transition, spreading out the textural change over a longer period to build anticipation for the final form. next is the Static Crescendo. a normal crescendo, where the band slowly ramps up the intensity: (bang) is a great transition technique, but not really texture. it's more dynamics. or, ok, it's a little texture. everything is texture, but a static crescendo really plays with our texture expectations, taking a single, held sound and swelling up in volume. a common way to do that is what a friend of mine calls "preverb", where you take a sound with a long decaying tail, like a hi-hat, cymbal, or guitar chord, and play it backwards: (bang) so instead of trailing off, it sweeps back up to the level of the initial attack. this works basically the same as the melodic soar, but with dynamics instead of pitch. and finally, there's the Novelty Hit: this involves a new textural layer we haven't talked about yet, what Dr. Megan Lavengood calls the novelty layer. this layer is for really distinctive instruments that resist blending with the rest of the soundscape. it's mostly used for accents, so it's a natural way to set up a transition. (bang) it works kinda like the upbeat activity, but while that's usually foreshadowing some sort of timbre change in the new section, the novelty hit is a transient event, making its point more with its unique tone than its rhythm. after the hit, it usually goes away again, because sitting with a novelty sound for too long gives it time to integrate with the rest of the band. you risk losing track of the distinct timbre that makes it so effective in the first place. and as you've probably noticed in the examples throughout this video, these textural cues can also be combined, using multiple signposts at once to really direct your listener's ear where you want it to go. maybe you withhold the filler while the beat does a rhythmic acceleration or something. and, of course, there's one more way to use these cues: don't. a sudden, unprepared spike in textural power can also be really effective. (bang) there's also plenty of songs that just don't change texture very much between sections, and while you can still use texture cues there, you might not want to. no technique is right for every song in every style, but in the world of pop and rock, these cues can make a world of difference. one of the examples I used here was Ironic, by Alanis Morisette, and I've actually been thinking about that song a lot lately. it's one of those songs that gets made fun of for its admittedly kinda silly lyrics, which I think is a shame 'cause debating whether or not the thing she's describing is technically irony distracts from the fact that's it's actually a really interesting piece of music. and to prove it, I did a whole analysis video, which is up now on my Patreon. it'll go up on YouTube in, like, 3 weeks or so if everything goes to schedule, but I wouldn't be able to make these videos without the support of my patrons, so for a while now I've been uploading videos early over there to say thanks. they also get to vote on which songs I talk about, so it's their fault that I'm even looking at Ironic in the first place. if you want to see me think far too deeply about a kinda silly but also beautiful and poignant song, there's a link to my Patreon in the description! and as always, thanks for watching. thanks to our featured patrons, Susan Jones, Jill Sundgaard, Howard Levine, Warren Huart, Damien Fuller-Sutherland, Neil Moore, Geoff, and Michael Mol! check out Patreon for a fuller outro, and as always, keep on rockin'!