I am teaching two new NYU classes this fall. One of them is History of 20th Century American Popular Music, which I wrote about recently. The other is Pop Music Analysis for Non-Majors. I enjoy teaching new classes, even though they require a lot more work than teaching an existing class. The history class takes significant prep on my part, but it isn’t difficult work, because I know why people are taking the class and what we’re expected to teach them: in addition to factual information about the evolution of jazz and rock and so on, the students need to learn how to do research and how to present it in writing. By contrast, the pop analysis class should be a breeze to prep, because I have taught the subject matter a thousand times. However, it actually poses the bigger problem, because I am not as clear on what the learning goals of the class are.
Here’s the catalog description:
This course introduces students to music analysis through close listening to popular music. Students develop the vocabulary and skills to describe, analyze, and interpret harmony, melody, rhythm, form, and timbre, using a variety of online music theory tools and AI applications for composition and analysis. Combining creative practice with critical listening, the course offers a hands-on, technology-driven approach to understanding how popular songs are constructed and how they communicate musical ideas. Intended for non-majors, no prior musical training is required.
Based on the class list, I am going to have a diverse group of students to work with; their majors run the gamut from math to theater. I don’t know about their prior music background yet, but it’s safe to assume that it ranges from “accomplished classical musician who is curious about pop” to “self-taught singer-songwriter or producer with a ton of experience but no formal vocabulary” to “never touched an instrument before.” It is possible to teach all of these people at once, but it takes some creativity.
Let’s break the course description into smaller pieces.
Students develop the vocabulary and skills to describe, analyze, and interpret harmony, melody, rhythm, form, and timbre…
The vocabulary and skills to describe, analyze, and interpret harmony and melody are otherwise known as music theory. There is a limit to how much music theory it’s possible to teach a complete novice in one semester. So my approach is probably going to be more like, there is this thing called music theory, I will talk you through as much as we have time for, you’ll retain what you retain, but hopefully you get a sense of what the big conceptual categories are so you know what to look up on Wikipedia later. I’ll come back to this below.
As for rhythm, there is not as much formal vocabulary to begin with. The best way to understand it is through practical experience. I am a big believer in drum programming as a rhythm pedagogy method. Learning the basics takes a matter of minutes, and it’s a skill that you can apply in many different musical situations. I’ll be using the Groove Pizza and DAW piano rolls. Beyond the practical value of being able to make your own beats, it’s also massively clarifying to understand the conventional role of each component of the drum kit: kicks on the downbeat, snares/claps on the backbeats, hi-hats on the eighth note or sixteenth note subdivisions. So, for example, if you are not sure what the tempo is, if the snares are falling on two and four, then you’re counting correctly.
Analyzing form is less about theory and more about learning the conventional terminology: intro, verse, prechorus, chorus, bridge, breakdown, drop, and so on. The best way to learn this is just to do it. Brian Jarvis’ Briformer web app is a great tool for diagramming song forms. I also like to have students use spreadsheets or pencil and paper for this. The important thing is to decide which parts of the song are separate sections and which are subsections of a bigger section, and to be able to count their length in bars. There is some subjectivity involved here! You very quickly discover that pop music is not as formulaic as you might think, that the most popular songs tend to have the weirdest and most ambiguous structures. One fun exercise I like to do is ask, which part of “Single Ladies (Put a Ring On It)” by Beyoncé is the chorus?
Timbre is barely theorized at all, but it’s extremely important in pop music, especially in the past few decades. It’s not that difficult to distinguish the basic instruments: guitar, piano, bass, drums, horns, strings. It takes more practice to distinguish between the various electronic keyboard instruments: electric piano, electric organ, Mellotron, analog synths. You can also learn to tell acoustic drums from drum machines or sampled breakbeats without too much trouble. But once we get into the computer era, suddenly everything is synthesized or sampled, and also heavily processed, so it’s impossible to tell what you’re hearing half the time. The best we can do is try to give broad descriptions: a lead, a pad, a bleep, a growl.
Why do people need to be able to analyze production choices and song structures in the first place, though? If you want to write songs or produce them, the answer is straightforward: the best way to learn to create is by imitating tracks that you like. Anybody with a computer and a DAW can produce any conceivable sound, so the challenge is to choose from among the infinite array of options. If you know that you are going for a Depeche Mode vibe, then you need to be able to analyze Depeche Mode tracks and understand what is happening in them: what are the chords, what are the beats, what are the instruments and how are they layered, what are the effects, how is everything mixed and panned. Even if you aren’t going to be producing tracks yourself, if you’re a musician or singer you want to know how recording and production work, so you know how to communicate with producers.
But what if you aren’t a musician, a beatmaker or a songwriter, current or aspiring? What if you’re taking the class out of general curiosity? I certainly think it’s worthwhile for any music listener to learn how to listen analytically and critically, but it’s less obvious what they should be learning in a one semester class. It would probably be more valuable to be thinking about music history, about social and cultural context, about semiotics. The songwriters and producers need that material too; it’s just a matter of emphasis. So, a producer might need to know that the Beatles used a Mellotron flute on “Strawberry Fields Forever”, and how it differs sonically from an actual flute; a general listener should be more focused on why the Beatles used the Mellotron and what influence that choice had.
Whatever it is that people need to learn, I plan to spend some quality time walking through the multitrack stems of some iconic recordings: a Beatles song, a Michael Jackson song, a Depeche Mode song. I do this in every class I teach. We talk about the production, the arrangement, the mixing, the recording process, the chords, the scales, the form, the groove, and anything else that arises. There are not many multitracks in circulation and even fewer are available legally. You could do a similar process with stem separation, I guess, but there is no substitute for the originals.
Anyway, let’s continue on with the course description.
…using a variety of online music theory tools and AI applications for composition and analysis.
Online music theory tools might include Musictheory.net, Teoria, or Ableton’s Learning Music site. I will also be steering the students toward more enjoyable interactive tools like the aQWERTYon and the Epic Online Orchestra. We will also be experimenting with the spectrogram in the Chrome Music Lab. People who want to get deeper into the weeds can try Notenscheibe by Paul Bretschneider. Guitarists can try out Mike Hadlow’s Guitar Dashboard. There are a million terrible vibe-coded circle of fifths visualizers cluttering up the internet as well; I guess I can teach folks how to avoid those.
AI applications for composition and analysis is where we start getting into difficult territory. “AI applications for composition” probably means Suno and Udio, right? I guess we can talk about those tools and how much I hate them. I think everyone should sign up for a free account with one or both and spend an hour messing around, so you can experience the joyless vacuum at their heart for yourself. I have heard some other well-meaning educators talk about ways to use generative AI for critical listening and such, and I guess I’m open to it. Actually that’s a lie, I’m not open to it at all, except insofar as it helps us identify AI slop and articulate its harmfulness.
As for AI applications for analysis, I am in favor of students using stem separation as a music learning tool. It’s not easy to learn to pick out, say, the bass from a densely produced track. If you can hear the bass by itself, then when you go back to the full track, it is much easier to hear it there too. I prefer using original multitrack stems for this purpose, because stem separation doesn’t always work well, but multitracks are rarely available and stem separation is better than nothing.
There are also various AI transcription tools, chord identification tools, and so on. We should talk about these, because people should know they exist, that they kind of work okay, and that you should be very skeptical of them when it comes to edge cases or anything remotely complicated or unusual. And the kids should know that while AI tools are pretty good at detecting pitches and so-so at detecting rhythms, they are not adequate at representing that information in readable sheet music. Writing out charts is going to be a human skill for the foreseeable future.
Finishing out the catalog description:
Combining creative practice with critical listening, the course offers a hands-on, technology-driven approach to understanding how popular songs are constructed and how they communicate musical ideas.
“Creative practice” means making beats, doing remixes, and eventually writing original songs. I’d be happy to approach the entire class as a songwriting/beatmaking class, because it’s easy to fold theory and analysis into those kinds of projects, but that isn’t quite what the department has in mind. If left to my own devices, I would say that beginner musicians should not be learning formal music theory at all, at least not as a primary goal. Students should just make beats and write songs, proceeding by intuition and rules of thumb at first, and I should supply theory concepts as they arise in the course of solving specific problems. If you are going on a long musical journey, then it’s fine to say, learn this abstract information, you might not need it now but you will eventually. But for non-majors, there is no later.
Maybe the idea behind doing theory in this class is that learning theory has value for its own sake, independent of any practical use it may or may not have. Some people do want to study music theory as an end unto itself. And that’s great, music theory is fun and interesting and anyone who wants to learn it should be able to for whatever reason. But what should they learn, and how much should we expect them to retain?
One idea I had is to go up to the meta level and spend the semester talking about what music theory is, why the version of it you got in high school seemed so unrelated to anything you were interested in, how it’s practiced and understood by jazz or rock or hip-hop musicians, and what is going on in the current academic literature. My assumption with this approach is that if you need to know the notes in the Db major scale, you can just look it up; what I can do is talk about why you would want that information, why it’s widely believed to be prerequisite knowledge for music making, to what extent that is true, and what are some other pieces of adjacent knowledge that you might want instead. For example, maybe you actually want Db Mixolydian, or Db major pentatonic.
With all of the above in mind, here are the projects I’m planning for this semester.
Prince bassline
“Kiss” by Prince (1986) has no bassline. You will add one. It should spell out the chords and harmonic rhythm, and should complement the melody without duplicating it.
I will supply the chord tones of the A7, D7 and E7 chords and show where the chord changes are, but it is up to the students to deploy the notes in a musical-sounding way.
String arranging
Add violin, viola and cello parts to “All Is Full Of Love” (Radio Strings Mix) by Björk (1997). Your parts should complement the melody and bassline without doubling them, and should mostly use chord tones.
Again, I will supply the chord tones; this is an opportunity to get a taste of arrangement and voice leading.
Miles Davis in black and white
Improvise or compose a solo on “So What” by Miles Davis (1959). Use the white keys of the piano for the A sections and the black keys of the piano for the B sections.
I’m taking advantage of the fact that the white keys play D Dorian mode, and the black keys play five of the seven notes in Eb Dorian mode. The point of this exercise is two-fold: to learn to hear and count through the A and B sections of the tune, and to get a taste of jazz improvisation and melody writing.
Musical space and structure graph
Choose a recording with at least four distinct musical and/or vocal layers and create a musical structure graph. List each song section with its start time and length in measures. For each section, list each audible vocal part, instrumental track, or other sonic element. Describe the sound and timbre of each layer as specifically as possible, including any audible processing and effects. Also describe its location in perceived musical space (far/near, left/right.) You can do this as a text document, a spreadsheet, or whatever other format makes sense.
I did this as a music technology grad student. It’s a very difficult and time-consuming exercise, but it’s revelatory. Even if you pick a song that you have heard a thousand times, listening this closely to so many different aspects of the song always uncovers new layers that you hadn’t noticed before.
Create a new instrumental for an existing acapella
I might do this twice, once with a formally straightforward and conventional song like “Call Me Maybe” by Carly Rae Jepsen, and once with a more formally complex and unpredictable song like “Starman” by David Bowie. I will definitely be supplying my sassy remixes for inspiration.
Create a top-line melody for an existing instrumental
Write an original melody from scratch on top of an instrumental. You may also write and sing lyrics.
For this one, I will probably use one of my tracks, or something by a friend or former student, so the class won’t be familiar with it and won’t have to fight their existing associations. I will give guidance on a scale or mode to use, but will encourage the kids to use some chromatic notes too.
So, that’s the plan. I will be adjusting this plan depending on who is in the class, what they know already, and what they want to find out. I haven’t decided whether to assign reading, or, if so, what to assign; again, that will depend who is in the class. It’s very likely that people will want to talk about copyright, sampling, plagiarism and related issues; students always do, and I am happy to accommodate them. They might want to go deeper on the technical side: how to use a microphone, how to use compression and EQ. I’ll be writing updates as we go.