Despite serious reservations about the merits of transcribing my improvisations, I have nevertheless done so, off and on, for forty years. These days, because I usually record to a MIDI file, I am able to easily access all the notes that I played and I have a good picture of the music in what’s known as a “piano roll” representation on the computer. But because I do not record with a click track, the notational algorithms of the various recording programs are rendered nearly useless. However, I have been transcribing my improvisations since before there were computers to capture them. My early efforts at transcription, from reel-to-reel tape recorders at half speed, taught me many things about both music notation and myself that I don’t think I would have learned otherwise.
The first thing I realized was that nearly all music that is notated was conceived that way. Composers are taught to write music deductively; that is, they are taught to derive more music from smaller bits of music whether it is motives, sets, sounds, or whatever. This process is carried out on paper (or more often on computer these days) and is done using music that is already notated. Even though music notation has been greatly expanded and enhanced these days with graphs and other formats, it is still an expanded use and development of ideas that are already in written form. Music is also taught in this written form, and so composers naturally have been conceiving of music in this way for hundreds of years.
I began to understand that, like mathematics, rhythm and pulse, though very useful tools, were not real in themselves.
When I first started to improvise, I used the ideas that I knew. I played in keys, meters, tempi, etc., but soon began to realize that that wasn’t really necessary. I found I didn’t need to think in a meter, I could think in phrases. I used a concept I called “macrometer” in which I thought of the focal point of the phrase as the “downbeat” and the rest of the music as leading either toward or away from that point. When I tried to transcribe these pieces, I found that putting the music back into traditional metric structures was difficult. Not only was it difficult, but it also could be done a number of different ways! I found that meter was a very fluid concept and not the musical foundation I thought it was. I began to understand that, like mathematics, rhythm and pulse, though very useful tools, were not real in themselves. They were concepts, but do not actually exist outside of the music that defines them. The pulse is defined by the music and not the other way around. This has been a very helpful concept to understand in ensemble playing as well. Members of an ensemble are not following the pulse, they are following each other! Collectively they are defining the pulse and the rhythm, not the other way around.
Tempo is also quite fluid. While transcribing improvisations, I found metric modulations, rubati, multiple tempo referencing (referring back to previous tempi) all turn out to be common and make the music quite complicated to notate. Because I conceived of the music using phrases instead of meter, tempo rubati happen often and are difficult to deal with. I still have not worked out a very good way to transcribe rubati. Recently, I tried transcribing all the rhythm against a fixed pulse so that different tempi were notated as tuplets or different note lengths. The method is more accurate but it is complicated, and the internal logic of the music becomes obscured. The difficulty of transcribing abstract rhythms and tempi has been a continuing problem and has caused me to return to the drawing board many times. I am beginning to think maybe I should be looking somewhere else.
I have heard conductors tell me to play exactly what the composer wrote my whole life, but a computer really does play what is written and it sounds terrible.
When I started using a computer and used a software program to play back my notation, I began to realize how very generalized music notation really is. I have heard conductors tell me to play exactly what the composer wrote my whole life, but a computer really does play what is written and it sounds terrible. There is no tempo inflection, no dynamic shading, no phrasing, no balancing, no shading or coloring, nothing other than the notes on the page, some articulations, and six or seven volume levels. To get the program to play my written notation so that it sounded like my improvisation took so many tempo markings, articulations, and dynamics that the music became hopelessly complicated and over-marked. I then would try to back off and leave some of them out, but where do I start? Classical music performers, especially pianists, interpret music by adding these things to the more generalized representation that a traditional composer writes. It is very difficult to try to make a recorded performance less specific and more generalized! I must try to imagine how much notation a performer would need to make what I actually did as an improvisation a legitimate interpretation. This is not music – it’s psychology!
Some of the most interesting things I learned from doing transcriptions were about structure, musical material, and memory. It is perhaps not very helpful to say that improvisational structure is “organic,” but that is the best word for it. The way the music is organized is not hierarchical but “linked.” New phrases both react to previous phrases and borrow material from them. Musical ideas seem to hang around for several phrases and then give way to newer ideas. Sometimes older ideas would be referenced, but not regularly. A new idea would be like throwing a stone in the water; it would ripple strongly for a few phrases and then peter out, giving way to newer material. There often seemed to be a subconscious attempt to balance new material with old material, though that pendulum would swing in either direction from time to time. “Organic” in this sense means a gradual unfolding of new material mixed with and as a reaction to older material.
Before I went all in on improvisation, I went through a period in which I would use an improvisation to craft a larger composition. What I tended to do with my improvised phrases, when I used them in my composition, was to stretch or extend them into larger phrases or sections. In the meantime, when I wasn’t composing, I would continue to improvise. However, the work I was doing in my composition slowly worked its way into my improvisation to the point where I no longer felt that lengthening or extending the phrases was necessary. It was at this point that my improvisation became my composition.
However, this “organic” structure was not just one continual process. There were sections and larger sections that were identified by texture, cadential syntax, or other distinctive changes. I never am aware of the larger structure while improvising, though I am aware of the changes, which I often identify as opportunities. I do often have a certain “feeling” about where I am within a structure, somewhat analogous to the feeling of direction you get when hiking in a forest. Of course, sometimes I end up wandering in circles, but I don’t transcribe or even save those improvisations!
When you are composing and building your musical structure with motives or other types of material, that building-block material doesn’t change. Any alteration that is done would be inferred as deliberate, which implies “with good reason.” Even if you did decide to “transform” the material, it almost always would remain recognizable. There is little reason to not do it that way. When you are improvising, however, you must rely upon your memory, and memory is fickle and impressionable, especially when distracted by the constant act of creation. What you remember from what you have played may not be its most distinctive characteristic, or maybe you don’t remember all of it. You may get the intervals right but not the rhythm, or vice versa. And when improvising, the music is much more likely to be altered by context, not to mention an occasional accident.
As a result, the musical ideas change. The ideas that do appear seem to “evolve” out of what comes before. Sometimes, referencing that material is identifiable, but sometimes it is not. It is a little like the childhood game of Pass The Secret, where you whisper something to the next person and they to the next until it returns to the original person completely changed. [A fun variation on this game is to copy a phrase or story (a joke is especially good) into a translation program and take it through several languages before translating it back to the original language!) When we retell or reference something, we often tell it in a different way. In music, I have come to the point where I think this type of development is more natural. The music grows, instead of being constructed from identical musical bricks. To me this feels more honest, while constructed music feels somewhat affected. Nowadays, I grow weary when I hear composers using the same material all the time. When you do a lot of improvising, composed music sometimes begins to sound rather “contrived.” Such is life.