How I learned to write music in real time
And how you can too, maybe
For as long as I can remember, I’ve known in advance of pressing a piano key, exactly what it was going to sound like. I’m not quite sure where I got the ability from — likely in part because we had a piano in the house since I was very young. Mum used to play a little, and she listened to lots of Mozart.
Perfect pitch is by no means essential to writing music, improvising, or even playing by ear, so don’t worry if you want to do these things and you don’t have it. But a good scientist tries to acknowledge their own biases, and in this case perfect pitch is one of mine.
As I showed interest in playing, my parents sent me to piano lessons. I’m eternally grateful for these, because even though they turned practice into a chore which I hated, they developed my skills enormously. I worked my way up through the Associated Board grades, but not all the way to the top, because at some point I realized they weren’t teaching me what I wanted to learn.
I wanted to be in a band, because obviously that would be cool. Reciting Beethoven wasn’t going to cut it, and in any case any CD player could recite Beethoven better. I wanted to create my own material, accompany others, play by ear, and jam — improvise. There seemed to be plenty of people in the world who could do these things, but nobody who really taught them (I did ask around a bit).
So I learned about chord notation. It’s not hard for a logical thinker: majors, minors, add2, sus4, 7th, maj7, 9th, 11th, augmented, diminished, etc. But it wasn’t a purely theoretical exercise; I always stopped to ask myself what each chord actually felt like. To play songs by ear I tried to recognise the chords first, and figure out how to play them after. I spent ages trying Bridge Over Troubled Water, but in the end gave up and had to buy the piano score. No matter, because when I learned music from a score, I wasn’t just learning the music, I was forming an abstraction for the techniques it used. Improvisation is vastly helped by having a large repertoire of tricks to choose from. Bridge Over Troubled Water, as it turned out, taught me a new way to play chords, melodies and bass simultaneously, which is always a challenge when you only have two hands.
As I learned about chords, I also learned about what combinations of them I preferred. I’ve always been a visual thinker, and I would draw huge maps that linked each chord to the places I thought sounded good next (directed graphs, in computing terms). Of course these things weren’t comprehensive, but they were a fun exercise. I still remember one of my favourite cycles from one of them: C, Cmaj7, C7, F, Dm, Dmaj7, D7, Gsus4. Yeah, it’s pretty cheesy; I’ve always struggled with that. But then I’ve been to some godawful performances by people who try to eschew all forms of harmony, melody and rhythm altogether, and that’s not my thing either. Talking of cheesy, at some point my parents bought me a Yamaha PSR400 keyboard, which was great for exploring chord patterns using the built-in accompaniment feature. The cheap sounds and the fact I didn’t have to think about rhythm probably held me back in some ways, but it was a great tool for learning about harmony, melody and song structure.
(As an aside, what does “cheesy” actually mean? I would argue that music is fundamentally about manipulating the listener’s emotions, with their willing consent of course. Cheesiness is where the listener becomes too consciously aware of the tricks you’re using to do this. So yes, that makes cheese subjective; different people will find different material cheesy).
Anyway, yes, chords. The above list doesn’t quite cover everything, because you can put different notes in the bass, for example C with G in the bass is notated C/G. I went through all these combinations trying to learn, to really internalize how they felt. In computing terms this adds another whole dimension to the dataset. And more dimensions start stacking up: the different inversions you can use for the treble notes, for example. Then there are feelings you get from transitions between them. Play C then C/E. How would you describe it? A feeling of stepping forward? Maybe it’s different for you. And you can start to analyse melodies like that too, in terms of the jumps between pairs of notes, then threes. Big jump up followed by small jump down, gives a very different feel to big jump up followed by small jump up.
Once you learn enough chords you realise that many are ambiguous. Is it C6 or Am/C? On one level it doesn’t matter, it’s the same sound. But the notation had given me two different ways to think about it, so it had actually enriched my world. At some point I started cutting notation out of the loop altogether, and just connecting sounds I wanted to hear to other sounds I wanted to hear, and connecting sounds directly to movements of my fingers, without the intervening notation. This was effectively returning to how I started as a child, but this time I was a lot better.
Jimi Hendrix managed to skip the notation phase altogether, because he never learned to read music. I wonder whether that would have worked for me? Perhaps, but I might not have become very technically proficient without the challenge of reciting classical music. Playing other people’s notes always forces me out of my comfort zone — breaks the loop of improvisations that have become more habit than creation. It’s no coincidence that the hardest gig I’ve ever had was playing music from a score (Claire de Lune, at my sister’s wedding). Anyway, though I no longer think in terms of western notation while playing, my perception of music has been forever shaped by it. I don’t think that’s either good or bad, it just is.
You could also frame the issue of notation (or lack of it) as a creative tool, as conscious versus subconscious brain function. It’s interesting that some of the music I’m most proud of I wrote while asleep. I have often had trouble sleeping; at one particular time I had an annoying habit of jolting awake only a moment after I’d dropped off. There was a silver lining however, because in the brief dreams I had I would hear some of the best sounds I’ve ever had the pleasure of knowing; like I’d encountered raw Platonic ideals for concepts like uplifting or funky. On waking I would try to write them down (notation is useful for that at least) or sometimes wake my partner in an effort to whisper cryptic notes into a voice recording app. The tunes I wrote as a result didn’t quite live up to the dreams, but I still think the riffs are the best I’ve written (funky one here).
Back to the added dimensions of tonal space — basses, inversions, movement, etc. At some point there were just too many to explore, so rather than explore them all, I started to form generalizations. In computing terms, using dimensionality reduction rather than exhaustive queries on the data cube. Do we want more harmony or more dissonance? Is this chord dense (like 2nds, 4ths etc) or sparse (like 5ths, octaves)? More notes or fewer, Herr Mozart?
Another angle into all this was jazz theory. In 1997 I saw the World Wide Web for the first time, and found a site called Mark Sabatella’s Jazz Improvisation Primer. I never really listened to much jazz, but I learned that you could think of melodies in terms of the scales they were written in: the classical modes; Dorian, Phrygian etc, open scales, bebop scales (which conveniently add an extra note so the scale length synchronises with the bar length — check out the solo from this live version of Walk of Life) and of course blues scales, which have the property of sounding great no matter how many mistakes you make.
When I was 17 I was given an old Macintosh SE by the head of IT at my school (Andrew Walwyn: many found him odd, I found him inspirational). I soldered together a MIDI interface, connected it to my PSR400, and started sequencing MIDI with an old copy of Digital Performer. I wrote a lot of my own material, little of which was worth saving. But I was also continuing to learn by copying sounds I liked. I reproduced “He’s On the Phone” (St Etienne) by ear and got it as near to perfect as you could on a PSR400, I think. It had lots of open intervals, which I found very emotional; I think most people do. Why is that? I have a pet theory that the dip in the frequency spectrum, e.g. between the G and D in a C9, reminds us of the resonance our own head has when we have a tight throat, are upset or crying.
Analysing music I loved was a key step towards writing my own. I would pause the best bit of the song, rewind, play again. Why was that bit so good? Try to generalize, abstract.
Then there was accompaniment, which is a funny skill. You don’t want to drown out what the other person is doing, so you end up playing different rhythms altogether; often syncopated, off the beat. The most intense part of your performance should probably fall where there’s a lull in the other person’s. The whole thing is a mirror image of the main tune, a polar opposite. It took me a long time to learn this; my development of rhythmic skills has been far behind harmony and melody.
Can all of this be taught? I think it can, but other than sharing my own journey, I’m still not quite sure how. I’ve put a couple of links below.
Is perfect pitch a help or a hindrance? Don’t get me wrong, I love it, but I’m certainly below average in my ability to transpose music — even just singing — because to me, each key has a different and unique texture. I fear when I write music that I’m making artistic decisions based on differences that are irrelevant to most people — even another listener with perfect pitch would probably perceive different textures to mine. But if I don’t write for myself, how can I expect anyone else to like it? My favourite key is E minor; if I don’t consciously choose not to, I tend to write in E.
In any case my perfect pitch is no longer quite perfect; sometimes it drifts off by a semitone. I put this down to two things. One is the onset of hearing loss. The other (likely not a coincidence) is listening to lots of electronic music, often mixed by a DJ who will have pitched tunes up or down to match tempo. So each tune is no longer in a key close to the frequency series 440*2^(n/12) Hz. I’ve lost my reference points, but I’m not so bothered about that because in the process I listened to lots of excellent music.
World Renowned Piano Pieces contains lots of technical challenge in one volume including Debussy’s Claire de Lune.
The piano score to Bridge Over Troubled Water employs good technique for simultaneous chord + melody playing. Don’t get one of the ‘easy’ versions.
Keyboards: They don’t make the PSR 400 keyboard any more, but the PSR range is still going strong!
Improvisation and the Theatre by Keith Johnstone, I am currently reading, in fact it inspired me to write this post. Lots of interesting things to say about improv.
Improvisation at the Piano I bought with a view to teaching, and it does look to have some useful exercises.
Here’s Mark Sabatella’s Jazz Improvisation Primer, free pdf :-)