This guide is being written in response to a question from one of my friends at CSUN, and I thought it might be useful to post it more generally as well. He’s working on a large-scale musical project right now and wanted to know how to go about starting the musical work. We discussed some aspects of this as film scoring majors at Berklee College of Music, because feature film scores in particular will also need to be planned on a large scale, and sometimes similar ideas can be useful in other multi-movement contexts as well.

So how should one go about writing a large scale musical work? The most obvious way of doing this that occurs to most people early on is to start writing at the beginning and then work forward from there. This is the way in which the audience will experience such a project in the end, but it is usually not the best way to approach a project from the creation side. It works for relatively small scale projects of only a few minutes, but tends to lead people to bog down - I can’t tell you how many “beginnings” of projects I have on my computer, where I had a cool idea to start a piece and then couldn’t figure out what to do next. They are sitting there, waiting for the day that I come back and pick them up to do something, which does occasionally happen - my piece “Overture for the Planet Earth” for orchestra sat there for about 10 years between coming up with the opening theme and writing a finished piece that could be performed by an actual orchestra. But even that is a relatively small-scale project, a single movement that lasts about 7 minutes. When organizing a piece like a full film score, or an opera (which is what my friend is working on now), it helps to have a more coherent plan from the beginning.

At Berklee, we discussed planning a project with a “concept”, which is an actual written document outlining the nature of the proposed score. Film will be different from something like an opera, because usually with a film score, you are starting with a hopefully more or less finished picture / dialog etc, and then adding music to an existing project; whereas with an opera, the music comes if not first, at least early in the process and the visual elements are made around the existing music. You would start with a libretto, basically the script, and then add music around that before doing production work on sets and blocking with actors, etc. So they aren’t the same, but there are still some related ideas between them.

The concept document will discuss the nature of the music and what elements of the existing story you want to reflect in the music. The example I remember from our class at Berklee was a movie that was set in Ireland. You may decide that you want to write a Celtic-influenced score because of the setting, and that can be a perfectly valid thing to do in this context if the setting is important to the story. But if the story just happens to be set in Ireland and that is incidental to a story that is more a gritty coming-of-age story in (say) Northern Ireland in the time of the Troubles, then perhaps using a Celtic score may be less relevant. That will be up to the composer to work out in conjunction with the director and producer on the project. However you do it, the concept document could address some of the following elements:

Who, What, When, Where, Why - 

Who are the main characters and what kind of people are they?

What is the overall plot of the story? Look for the arc of the narrative.

When does it take place? Period dramas often use music that reflects the time period, though again, you’d want to consider the degree to which the period was a crucial part of the overall story or whether it was more universal in nature and just happened to be set in that period.

Where does it take place? See above for a brief discussion of this point.

Why - what is the overall point of the story? Is there a particular moral or “thesis” that they explore through the story? Sometimes they hit you over the head with these elements, other times they are much more subtle (I tend to like that approach better, but it depends on who the target audience is and if there is another point to the production). 

Also, consider “why?” to be about justifying the musical choices made in the other points as well.

You will also include more specifically musical aspects in the musical concept document - based on the analysis you provided above, what sort of instruments would make sense for this project? Should it use standard orchestra? A rock score? Orchestra + ethnic instruments from a particular culture? Maybe an entirely world / ethnic score? You want to have a coherent idea for the overall score, which can be modified with specific reason for particular scenes, but you don’t want every scene to be taken in isolation and wind up separated from the overall project, as this will lack cohesion and lead to a broken up feel for the overall project. At that point, they might as well score the project from a production music library and save the money on hiring a custom composer (obviously that applies more to film than opera, but operas should be made coherent too).

With opera, you will probably start with a decent idea of the nature of the score from the beginning, but you can write this up for that kind of project as well to be sure that it all makes sense on a conceptual level.

The other part of the concept will be describing the idea behind what the music will do. This will vary more from scene to scene and be described in more detail in the next phase, but you can mention broad ideas in the concept - is it a primarily psychological score, revealing the thoughts the characters don’t say out loud (this is one way to show irony in a scene where it may not be super apparent otherwise, for example). Is it primarily a background element not really “doing” anything in particular? Maybe it’s even “Mickey-Mousing”, catching as many elements happening on screen as possible. This kind of music is very busy, can be complicated to write and record, and is often more over-the-top than is really required, but in the right context it can still be effective. It’s more often used in animation than in live action contexts, hence the name, and I’ve never really seen it used in a stage production.

Once you have an overall concept, then it’s time to begin getting more specific. In film scoring, the next part of the process is called “spotting” the film, and we create “spotting notes”. There is a very specific format that these usually take (read more about it here), but the ultimate goal is to have a single document that lists every element of music that will be used in the project. We call these “cues” in film music, which can sort of be thought of like movements in a standalone piece of music, though many times cues will be very short. We want to give each cue a “Cue Number”, which will take the form of #m# (1m1, 1m2, 2m1, etc). The first number usually refers to a large grouping - originally the reel of film the cue happened in, nowadays it might be an “act” or a single hour of a multi-hour feature film, or however you choose to organize it. Some projects just use “1” as the first number for all cues today. The “m” means music (as opposed to audio for dialog or sound effects, etc), and then the last number is the order the cue appears in relative to other cues in that overall group. A feature film can very well have 40+ cues in it, while a short film might still have 10.

Some cues might be “source cues”, where the music actually appears from a visible source on screen - the radio in a car, a band playing in a club, a string quartet playing at a formal event, etc. One key point about source cues is that the character(s) in the scene can also hear the music. Generally characters will not hear or react to the underscore cues except in certain comedy situations where they “break the fourth wall” and make the characters reveal that they know they are characters in a film or show. The cue numbers for source cues will have “S” appended to them - 3m2S, for example. These will usually be handled through licensing deals with the music supervisor for the project rather than by the composer. My day job for the past decade-plus has been running a production music library, and many of our library placements were for source cues in film and tv shows.

At the spotting session, the director, producer, and composer (and anyone else who has a controlling interest in the music) will go through the film together and figure out the exact frame that each cue will start and stop on, and this will be included in the spotting notes using SMPTE (pronounced “simpty”) time code. This takes the form of HH:MM:SS:FF - hours, minutes, seconds, frames. Sometimes hours will be reel number instead, but these days it’s usually in hours. The start and stop codes will be given, which will allow the exact length of each cue to be calculated, and then the exact length of the full score can be calculated from that information.

Once the locations for each cue have been chosen, then more detailed information about each cue can be provided - what the musical function of each cue is (see above for more on that), any notes about specific instruments that would be needed, certain actions to hit (make that screen action line up with a beat in the music), etc. By the end of the spotting session the composer should have a pretty good idea of the scope of the project and can then go and begin the process of writing the score itself.

All of this information about spotting notes really applies specifically to film and tv, opera would work differently. Again, generally the composer will know more of this kind of information in advance on an opera score, since the composer often has more direct creative control of the overall project when writing an opera (particularly if they also write the libretto - if they work with a librettist then they will need to coordinate with them for some of these elements, and if it’s a commissioned opera, then certain elements might be dictated by the commissioning institution).

For both kinds of projects, the next step I’d consider is any musical themes that will be used in the project. We often think of themes in a melodic sense - tunes that are used throughout a project to invoke various ideas or characters. Richard Wagner was the first major composer to make use of a technique called “Leitmotif”, which were musical elements in the score that were supposed to directly invoke the various characters and places, etc. He could then develop the musical ideas to reflect the development of those elements in the story. If someone was upset, their theme might be played in a minor key, if they just won something, it might be played triumphantly, etc. This can also be a way to reflect aspects of the story that are difficult to show on stage or on screen, when a character is conflicted about something, or to set up foreshadowing, for example. This idea of Leitmotif was adopted by other opera composers after Wagner’s time, and then became a common concept in film music when that grew out of opera scoring in the early 20th century. Star Wars would be a great example of a relatively recent film that made heavy use of leitmotif, as John Williams pulled a lot of his musical ideas out of classical opera and stage music, as well as from film scores of the 1930s that themselves pulled heavily from that tradition. These days, themes don’t have to be melodic in nature. Jeff Rona in The Reel World talks about using specific synthesizer sounds or unusual acoustic instruments in a thematic manner to reflect the villain in some of the projects he’s scored over the years.

Coming up with some sort of themes for a project would be a good next step. Many composers start by writing a suite, or an overture of sorts, to try out some of the thematic ideas they come up with and play with them in different dramatic contexts (how do you develop an idea if you need it to feel sad, happy, excited, depressed, etc). This piece of music could be the actual overture of an opera, or often finds its way into the credit music in a film, or sometimes it never sees the light of day outside their studio or their music team. These days, there’s a trend of releasing film music in a manner that it can be played in a live concert, and sometimes this suite is a good way of doing that too. But it’s a useful piece of music to have, both because it’s often one of the elements that feels most like straight-up concert composition, and it can guide the rest of the music team if there are other people involved in writing “additional music”.

From there, there are all sorts of ways to approach writing the cues. Sometimes you want to start with the cues that you can basically just “hear” already, before going back to work on the more difficult cues. Or maybe you want to start with the climax of the project, so you know what you are building to with the rest of the score. There are probably as many ways of doing this as there are composers, and maybe even as there are projects! But I would think it would be pretty rare to start with the first note the audience will hear and go straight through until you write the last note the audience will hear.

I will also note here that you should expect to do a lot of revision on a project like this. Particularly for film scoring, where you have to write music that is acceptable to a (usually demanding) client, you will almost certainly not nail a cue on the first attempt. On the projects I’ve been involved with as an assistant or copyist, I’ve more often seen version number of v3b than v1a. The scheme we used on those projects was v(ersion)[Major revision #][Minor revision letter]. So v3b would be the third major rewrite, and then the second tweaked version. The first rough draft would be v1a.

I once had a professor at Berklee who gave us two grades for our projects. The class was “Contemporary Techniques in Film Scoring”, which was basically about synthesis techniques as applied to scoring short scenes. Each project got a “class grade”, which was basically “did you do the assignment?”, and then a “real-world grade”, which was the notes a director might have had for you. It wasn’t hard to get an A for the projects in the class - one was about writing a piece where we used delay effects to create sounds that were distinctly different from the undelayed sound. If you did that, the grade book would record an “A”. So I’m sure many people got As in the class. But no one ever passed the “real-world” component - there was always something that he would want us to tweak or throw out entirely.

Several of my Berklee classes mentioned this point that the sort of experience we got in the classroom was in a significant way misaligned with what we’d need to do in the world. Namely, in real-life, you don’t get to walk away with a poor grade and move on to the next project. You can’t “settle” for a C and decide that’s good enough (you shouldn’t anyway, but that’s another story!). If you want to get paid for your work, you need to fix it if the director has relevant feedback. Our professors always said that it would be nice if we could try that in our classes, but the only class we really got that in was our final “Directed Studies” capstone class, where we spent several weeks scoring a short film with feedback from a private instructor from the department. More or less like the private composition lessons I get now at CSUN for my master’s degree work. My professor has feedback, and then I go home and tweak whatever he suggests how I see fit, and then bring it back the next lesson and see if he likes it better.

But a larger point that I’m making here is also not to feel bad if you do need to revise a cue quite a few times. Another point that Rona made in his book was that every time he saw a film with music he thought wasn’t what it could be, behind that was a director who said “Yes, that’s what I want”, and maybe even forced the composer to write that music against their better judgment. It’s a collaborative process, and while that can lead to better outcomes, sometimes it can also lead to worse outcomes, particularly when one person has overriding authority and they may not entirely know what they are doing.

So at any rate, hopefully some of these points are helpful when it comes to thinking about organizing a larger project. As you can see, when you have a large-scale project like a multi-cue film score or an opera, you need to do some pre-planning to make sure everything lines up correctly and nothing is missed. This also helps when you have several people involved in a production who need to be on the same page in their creative vision - in a film context, this would include the director, producer, and composer of course, but also the music editor, any production assistants for the composer, additional music composers, orchestrators, copyists, engineers on recording sessions, and also the engineers on the dubbing sessions where the final recorded and mixed music will be mixed in to the overall soundtrack for the film. There’s a reason that end credits in a film can last 10 minutes!